<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by DexSnipe on Medium]]></title>
        <description><![CDATA[Stories by DexSnipe on Medium]]></description>
        <link>https://medium.com/@dexsnipecom?source=rss-2654b58e3d1e------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*ayV9LZH3f4seVSlGuS0eTQ.png</url>
            <title>Stories by DexSnipe on Medium</title>
            <link>https://medium.com/@dexsnipecom?source=rss-2654b58e3d1e------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 02:14:24 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@dexsnipecom/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Solana Copy Trading Bot: How to Vet a Wallet Before You Copy It]]></title>
            <link>https://medium.com/@dexsnipecom/solana-copy-trading-bot-how-to-vet-a-wallet-before-you-copy-it-cdc5e833bfde?source=rss-2654b58e3d1e------2</link>
            <guid isPermaLink="false">https://medium.com/p/cdc5e833bfde</guid>
            <category><![CDATA[memecoins]]></category>
            <category><![CDATA[crypto-trading]]></category>
            <category><![CDATA[solana-blockchain]]></category>
            <category><![CDATA[trading-bot]]></category>
            <category><![CDATA[cryptocurrency]]></category>
            <dc:creator><![CDATA[DexSnipe]]></dc:creator>
            <pubDate>Thu, 19 Nov 2026 11:19:20 GMT</pubDate>
            <atom:updated>2026-11-19T11:19:20.123Z</atom:updated>
            <content:encoded><![CDATA[<h3>We ran the 15 busiest Pump.fun wallets through five simple checks.</h3><p>Most people pick a wallet to copy because it showed up on a hot coin and looked busy. We took the 15 busiest wallets on Pump.fun one morning this week and checked them the way you should before copying. None of them passed, and the reasons are more useful than the score.</p><figure><a href="https://www.dexsnipe.com/"><img alt="DexSnipe banner reading Copy the wallet, not its luck, with the address https://www.dexsnipe.com/ and an Open dexsnipe.com button. A dashboard illustration lists four snipers: Copy trade on 3 wallets at 0.1 SOL fixed that sells with them, Copy trade on 1 wallet at 25% of their buy that skips small buys, Creator watch on Pump.fun for the next launch from 1 wallet, and a Copy trade on 5 wallets on trial, switched off." src="https://cdn-images-1.medium.com/max/1024/1*vpifs4fweDar6Z5DvyPjHw.png" /></a><figcaption>DexSnipe is our own Solana sniper bot service.</figcaption></figure><h3>The short answer</h3><ul><li><strong>Busy is not the same as good.</strong> Of the 15 most active wallets in our sample, two never held a token of their own, eight bought 0.012 SOL or less at a time, and two only ever did half of a trade.</li><li><strong>Some wallets split the job in two.</strong> One wallet paid 11.6 SOL for 48 buys. In the same transactions, a second wallet received the tokens, and later it did all the selling. Copy either one and you get half a trade.</li><li><strong>Speed decides a lot.</strong> Two wallets held their median coin for 1 second and 5 seconds. A copy that lands a moment later buys after they are already out.</li><li><strong>Look past the best trade.</strong> The one slow, human-paced wallet in the group was up 0.86 SOL. Without its single best trade it was down 1.21 SOL.</li><li><strong>Five checks take about ten minutes per wallet</strong>, with a free block explorer. They will not find you a perfect wallet, but they will save you from most of the bad ones.</li></ul><h3>What a copy trading bot actually does</h3><p>A copy trading bot watches one or more wallets you choose. When one of them buys a token, the bot buys the same token from your wallet. When it sells, the bot can sell too. That is the whole idea, and it is a good one on paper: someone else does the research and you ride along.</p><p>On Solana the bot does not get a phone call when your wallet buys. It sees the trade on chain, the same way everyone else does, and then sends its own transaction. That gap is small. On new Pump.fun coins it is also expensive, because prices on a bonding curve move with every buy. The wallet you copy buys at one price. You buy after it, at a higher one.</p><p>Most copy bots let you set the same handful of things:</p><ul><li><strong>Which wallets to follow.</strong> Usually several at once.</li><li><strong>How much to buy.</strong> A fixed amount, or a share of what the wallet spent, with a cap.</li><li><strong>Which of their buys to skip.</strong> For example, anything under 0.5 SOL.</li><li><strong>Whether to sell when they sell.</strong> And what to do if they never do.</li><li><strong>Your own filters and exits.</strong> Checks on the coin before buying, and a take-profit or stop-loss that do not depend on the other wallet.</li></ul><p>The settings matter, and we come back to them near the end. The bigger decision comes first: which wallet you point the bot at.</p><h3>Why copying is harder than it looks</h3><p>Earlier this month we replayed 2,023 real wallet trades on new Pump.fun coins and asked a simple question. What if a copy had bought and sold two slots after each one, less than a second later?</p><p>The copy did worse than the wallet it followed 94% of the time. At the same 0.1 SOL per trade, the wallets averaged -10.5% and the copies -19.0%, before fees. Even a copy in the same block, right after the wallet, did worse every single time.</p><p>The gap was not the same for every wallet. It was smallest on wallets that bought a few minutes after launch, where 55% of the wallets’ trades were winners and 44% of the copies were. It was huge on wallets that bought in the first blocks. There, 55% winners for the wallets became 16% for the copies. If the wallet you copy is a launch sniper, you are mostly copying its entry price, and you will not get that price.</p><p>The same replay showed something else worth knowing. Copies that sold when the wallet sold averaged -19.0%. Copies that bought the same way but held until the end of the 30-minute window averaged -31.1%, and only 10% of them ended in profit. The wallet’s exit is half of what you are paying to copy.</p><p>We wrote up the timing side of that replay in <a href="https://medium.com/@dexsnipecom/how-to-snipe-meme-coins-on-solana-what-2-023-replayed-trades-say-about-timing-fees-and-exits-524c50d4e00e">how to snipe meme coins on Solana</a>. The short version for copy trading: pick wallets that do not depend on being first, and keep their sells.</p><figure><img alt="Chart of what a copy gives up, from 2,023 real Pump.fun wallet trades replayed with a copy two slots behind. Wallets that bought in the first blocks had 55% winning trades and their copies 16%. Wallets that bought minutes after launch had 55% and their copies 44%. At 0.1 SOL a trade the wallets averaged -10.5%, copies that sold with the wallet -19.0%, copies that held 30 minutes -31.1%, and 94% of copies did worse." src="https://cdn-images-1.medium.com/max/1024/1*TB9cB_f1VzxbTtlJ1DjlCw.png" /></figure><h3>How we picked the 15 wallets</h3><p>We wanted the wallets a new copy trader would actually stumble on: the names that keep popping up in the trade list of every fresh coin.</p><p>So on November 19, 2026, we pulled the last 3,026 successful transactions on the Pump.fun program, which covered just one minute, from 10:56:44 to 10:57:45 UTC. That alone tells you how busy Pump.fun is. We read 1,200 of those transactions in full and counted who signed them. 95 different wallets bought or sold in that minute. We took the 15 that appeared most often.</p><p>For each of the 15 we then pulled its latest signatures and read its last 150 transactions, or all of them if it had fewer. From those we worked out how often it trades, how much it spends, how long it holds, how many transactions fail, and what each finished trade made or lost in SOL, fees included.</p><p>This is a small sample from one morning. It is not a ranking of Pump.fun traders, and a different minute would give you different names. What it does show well is the kind of wallet you meet when you go looking, and why most of them make poor copy targets.</p><h3>Check 1: Is the wallet trading for itself?</h3><p>This sounds silly until you see it. The busiest wallet in our minute signed 1,000 transactions in under two minutes. That works out to more than 37,000 an hour. In the 150 we read, its own token balances never changed once. It paid the network fee, and the tokens landed somewhere else.</p><p>The second busiest did the same: 1,000 transactions in under 20 minutes, and not a single token of its own.</p><p>That pattern fits a wallet that signs and pays for other people’s trades, like the fee payer behind a trading app or a relayer. Whatever it is, it is not a trader. A copy bot following it would either do nothing, because no buy ever shows up as its own, or it would chase trades that belong to thousands of different users.</p><p><strong>How to check it:</strong> open the wallet on Solscan or any explorer and look at a few recent transactions. If the token changes on the transaction belong to other addresses and the wallet only pays the fee, skip it.</p><p><strong>Result:</strong> 2 of the 15 failed here. A third wallet, which we come back to in check 2, paid for its buys but never received the tokens either. That makes 3 out after the first check, and 12 left.</p><h3>Check 2: Does it buy and sell from the same wallet?</h3><p>This was the most interesting find of the morning.</p><p>Wallet A, ending in dLpx, showed up on our list because it sold a lot. In its whole history of 105 transactions, which covered less than two hours, it made 48 sells across 25 coins and took in about 13.3 SOL. On a quick look it was the most profitable wallet of the 15 by far.</p><p>But it never paid for a single token. All 48 times, tokens arrived in wallet A in a transaction where it spent next to nothing. Then, a little later, it sold them.</p><p>The money came from wallet B, ending in jw1U. Wallet B&#39;s whole history was 55 transactions over about the same two hours. 49 of those 55 transactions were the very same transactions that show up in wallet A&#39;s history. 48 of them were Pump.fun buys. In each one, wallet B paid the SOL, about 11.6 SOL across all 48, and the tokens landed in an account owned by wallet A.</p><p>So the two wallets were doing one job together. B bought. A held and sold.</p><figure><img alt="Diagram of one trade split across two wallets, read on chain November 19, 2026. Wallet B, 55 transactions in total, made 48 Pump.fun buys and paid about 11.6 SOL but kept no tokens. In the same transactions the tokens moved to wallet A, 105 transactions in total, which sold 48 times across 25 coins and took in about 13.3 SOL. Copying B copies every buy with no sell. Copying A shows no buy at all." src="https://cdn-images-1.medium.com/max/1024/1*ztp1ozmbVAHp7JHjVtubog.png" /></figure><p>Now think about what a copy bot sees.</p><ul><li><strong>If you copy wallet A,</strong> the bot never sees a buy. It sees tokens arrive and then sells. Your bot sits there doing nothing, or worse, it reads the incoming transfer as a buy signal and buys late into a coin that A is about to sell.</li><li><strong>If you copy wallet B,</strong> the bot copies every buy. Then it waits for B to sell, and B never does. The sell happens in wallet A, which you are not following. You are left holding the coin with no exit signal.</li></ul><p>Either way you get half a trade. It is the worst kind of copy target because the profit on the sell side looks so good.</p><p><strong>How to check it:</strong> look at the wallet’s buys. Does it pay SOL and receive the tokens in the same transaction? Then look at its sells. Do they sell tokens it bought itself? If tokens keep arriving for free, or if bought tokens keep leaving for another address, you are looking at one half of a pair.</p><p><strong>Result:</strong> wallet A failed here. 11 left.</p><h3>Check 3: Is the size worth copying?</h3><p>Eight of the 11 wallets left had a median buy of 0.012 SOL or less. Five of them sat at 0.002 SOL, one at about 0.005 SOL and two at 0.012 SOL.</p><p>They were also very busy. The 0.002 SOL wallets each signed 1,000 transactions in somewhere between 35 minutes and two and a half hours. In our 150-transaction windows they each made 27 to 41 round trips, in and out of the same coin, with a median hold of about 20 seconds.</p><p>And they lost, slowly. The five 0.002 SOL wallets were down between 0.001 and 0.004 SOL each across their round trips. That is roughly what they paid in fees. The pattern matches a volume bot, which buys and sells small amounts to keep a coin’s activity up. That is a service someone pays for, not a trading strategy, and there is nothing in it for a copier.</p><p>The other two small wallets, at 0.012 SOL a buy, were worse. One was down 0.02 SOL. The other was down 0.18 SOL over 47 round trips, with a median trade of -41.6%. That wallet’s entire history was 500 transactions in a little over two hours.</p><p>If you copy wallets like these at a fixed 0.1 SOL a trade, you are taking a bet ten or fifty times larger than theirs, on a coin they only touched to move a number. You also pay your own fee and tip on every one of those dozens of trades an hour.</p><p><strong>How to check it:</strong> look at the SOL spent on the wallet’s last 20 or so buys. If the typical buy is a few cents, it is not trading for profit in any way a copy can follow. We used 0.05 SOL as the line.</p><p><strong>Result:</strong> 8 failed here. 3 left.</p><h3>Check 4: Can your copy keep up?</h3><p>Two of the last three wallets were fast. Really fast.</p><p>The first held its median coin for 1 second. In 37 round trips it won 57% of the time and finished up 0.90 SOL, after paying 0.096 SOL in fees. It bought about 0.16 SOL a time. On paper, that is a nice wallet. In practice, a copy that arrives a slot or two later is buying while this wallet is already selling. The replay above showed what happens to copies of first-block buyers, and this wallet is even quicker than most of them.</p><p>The second held its median coin for 5 seconds and finished up 0.61 SOL over 10 round trips. Most of that came from one trade worth 0.46 SOL. It also signed close to 500 transactions an hour, and 87 of its last 150 transactions moved no token at all.</p><p>Neither is a bad trader. They are just not copyable. Their edge is speed, and speed is the one thing a copy cannot have.</p><p><strong>How to check it:</strong> for a handful of the wallet’s recent coins, note the time of the buy and the time of the sell. If most trades are over in seconds, skip it. We used 30 seconds as the line, and you can be stricter.</p><p><strong>Result:</strong> 2 failed here. 1 left.</p><figure><img alt="Funnel of the 15 busiest Pump.fun wallets in one minute on November 19, 2026, through five checks. 12 trade for themselves, after 2 only paid fees and 1 paid but kept no tokens. 11 buy and sell from one wallet. 3 buy 0.05 SOL or more, as 8 bought 0.012 SOL or less. 1 holds 30 seconds or more, as 2 held a typical coin 1 and 5 seconds. 0 profit without the best trade: +0.86 SOL became -1.21 SOL." src="https://cdn-images-1.medium.com/max/1024/1*f0e8VhcpgOuIH8AhRhnjpw.png" /></figure><h3>Check 5: Does it still make money without its best trade?</h3><p>The last wallet standing was the only one in the group that traded like a person.</p><p>It had only 297 transactions in its whole history, spread over 125 days. Our 150-transaction window ran from October 29 to November 19, three weeks of trading. In that window it touched 63 coins, made 94 buys and 53 sells, and still held 14 positions it had not sold. Its median buy was about 0.10 SOL and its median hold was 334 seconds, more than five minutes. A copy can keep up with that.</p><p>Across its 44 finished round trips it was up 0.86 SOL. That looks fine.</p><p>Then you sort the trades. Only 6 of the 44 were winners, 14%. The median trade lost 8.5%. One trade made 2.07 SOL. Take that one away and the other 43 lost 1.21 SOL together.</p><p>That is a lottery-ticket wallet: lots of small losses and the hope of one big hit. There is nothing wrong with trading that way. But copying it means you need to catch that one hit too. Miss it because your filters skipped the coin, your buy failed, or you were a few cents too late on a fast candle, and you are left with the 43 losses.</p><p><strong>How to check it:</strong> add up the wallet’s finished trades for the last couple of weeks, then do it again without the single best one. If the total flips from profit to loss, treat it as a lucky wallet until it proves otherwise.</p><p><strong>Result:</strong> the last wallet failed here. 0 of 15 passed all five checks.</p><h3>Two more signs worth a look</h3><p><strong>Wallet age.</strong> Three of our 15 wallets had a whole history of a little over two hours at most. One had 105 transactions, one had 55 and one had 500. Wallets that new have no record to judge. They are often made for one job and dropped after it. Wait a week and see if a wallet is still around before you trust it with your SOL.</p><p><strong>Failed transactions.</strong> Failure rates in our 15 ranged from 0% to 65%. The wallet at 65% had 652 failures in its last 1,000 transactions. A wallet that fails that often is usually fighting other bots for the same block. Your copy will be fighting the same fight, one step behind.</p><h3>So what does a wallet worth copying look like?</h3><p>We did not find one in this minute, and that is a fair result. Most of the wallets you see on a fresh coin are machines, and machines built for speed or volume do not copy well. But from the checks you can sketch the wallet you are looking for:</p><ul><li>It buys and sells for itself, from one wallet, with its own SOL.</li><li>It spends real amounts, 0.1 SOL or more per coin, so its trades mean something to it.</li><li>It trades at a human pace: a few dozen coins a week, not hundreds an hour.</li><li>It holds for minutes, not seconds, so a copy can arrive in time.</li><li>It is still in profit over a few weeks after you remove its best trade.</li><li>It has been around for weeks or months.</li></ul><p>Wallets like that exist. They are usually not the loudest names on a new coin’s trade list. You find them by following one interesting buyer back through their history, or by checking the early holders of coins that did well and seeing who keeps turning up, and then running the five checks before you add them.</p><p>One more thought. If a wallet you like is a creator, a wallet that launches its own coins, copy trading is the wrong tool. A creator’s first buy happens inside its own launch. You cannot copy that. A creator watch, where your bot buys the next coin that wallet launches, is a better fit. We looked at how fast the first buyers on a launch really are in <a href="https://medium.com/@dexsnipecom/solana-sniper-bot-who-really-buys-first-we-traced-411-early-pumpfun-buys-6cc602a342c2">who really buys first on Pump.fun</a>.</p><figure><img alt="Checklist to vet a wallet before you copy it: 1 does it trade for itself, its own token balances change; 2 same wallet for buy and sell, it pays for the tokens it later sells; 3 is the size worth copying, typical buy 0.05 SOL or more; 4 can your copy keep up, typical hold 30 seconds or more; 5 profit without its best trade over two weeks. Extras: weeks of history, not two hours, and a high fail rate means a block race." src="https://cdn-images-1.medium.com/max/1024/1*KpTBS6kM95VrhaqfpijDZA.png" /></figure><h3>Copy trade, creator watch or fresh launch: which one fits?</h3><p>Copy trading is one of several ways to point a sniper, and it is not always the right one. A quick way to choose:</p><ul><li><strong>Copy trade</strong> fits when you found a trader whose buys and sells you trust, and who is not racing for the first block. You get their picks and their exits, a moment later.</li><li><strong>Creator watch</strong> fits when the wallet you like launches coins instead of trading them. Your bot buys the next coin that wallet creates, in the first blocks, which a copy never could.</li><li><strong>Fresh launch</strong> fits when you do not trust any one wallet, but you do trust a set of rules: the dev’s share, the top holders, the mint authority, the starting liquidity. The bot buys every new coin that passes.</li></ul><p>Many people run more than one. A small fresh launch sniper with strict filters, a creator watch on two or three proven creators, and a copy sniper on one carefully checked trader cover different kinds of coins and rarely buy the same one. If they do overlap, a cap on how much you hold per coin keeps it from getting out of hand.</p><p>What all three share is that the bot is only as good as what you point it at. With copy trading that means the wallet. With creator watch it means the creator’s record. With fresh launch it means your filters.</p><h3>How to set up a copy sniper once you have a wallet</h3><p>Say you have a wallet that passed. Here is how we would set the bot.</p><ol><li><strong>Start small.</strong> Use a fixed size you are happy to lose, like 0.05 or 0.1 SOL, for the first week. You are testing the wallet and your settings, not making money yet.</li><li><strong>Or copy a share of their buy, with a cap.</strong> 25% of what they spend, capped at your fixed size, keeps your bets in proportion. When they buy small, you buy small.</li><li><strong>Skip their small buys.</strong> Set a minimum, like 0.5 SOL if the wallet usually spends a few SOL. Small buys are often tests or dust.</li><li><strong>Keep your own filters on.</strong> Your bot should still check the coin before it buys: mint and freeze authority, the top holders, the creator’s share. A good wallet still buys bad coins sometimes. Our <a href="https://medium.com/@dexsnipecom/pump-fun-sniper-bot-filters-we-checked-807-new-coins-to-see-which-rules-work-5c7a932f9e9e">filter test on 807 Pump.fun coins</a> covers which checks earned their keep.</li><li><strong>Sell when they sell.</strong> The replay showed this was worth a lot. Turn it on.</li><li><strong>Add your own exits anyway.</strong> A stop-loss and a timer protect you if the wallet holds forever or sells from somewhere you cannot see. A take-profit step lets you take something off the table if the coin jumps before they move.</li><li><strong>Set a fair priority fee and tip.</strong> You are not trying to beat the wallet. You are trying to land a slot or two after it, every time. Our <a href="https://medium.com/@dexsnipecom/best-sniper-bot-settings-for-pumpfun-slippage-priority-fee-and-tip-from-807-launches-76409e7a5040">settings write-up from 807 launches</a> has real numbers for slippage, priority fees and tips.</li><li><strong>Review after a week.</strong> Compare your copies to the wallet’s own trades. If you keep landing far behind or missing its winners, the problem is the wallet, the settings, or both.</li></ol><h3>Three mistakes we see with copy traders</h3><p><strong>Following the leaderboard.</strong> Leaderboards rank by profit, and profit over a short window is mostly one or two lucky trades. Run check 5 on anything you find on a leaderboard before you copy it.</p><p><strong>Copying too many wallets at once.</strong> Ten wallets at 0.1 SOL each can mean 0.5 SOL in open positions within a minute on a busy morning. Start with one to three wallets, watch how they behave together, and add more later.</p><p><strong>Turning off the sell.</strong> People turn off “sell when they sell” because they think they can hold for more. Sometimes that works. In our replay, holding after the wallet sold took the average result from -19.0% to -31.1%. The wallet usually knows something you do not.</p><h3>Where DexSnipe fits</h3><p>DexSnipe is our own Solana sniper bot. Copy trade is one of its six trigger types, and it is included in every plan. A copy sniper can follow up to 10 wallets. It buys at your fixed size, or at a share of the followed wallet’s buy capped at your size. It can skip their buys under a minimum you set, runs your filters before every buy, and can sell when they sell. Take-profit steps, a stop-loss, a trailing stop and a timer work alongside it.</p><p>Plans are flat: 1 SOL for 7 days, 3 SOL for 30 days or 5 SOL for 90 days. There is no fee per trade. You pay the network fee, your own priority fee and tip, and the venue’s trading fees.</p><p>What DexSnipe cannot do is make a bad wallet good, or make a copy as fast as the wallet it follows. The five checks above are still your job.</p><h3>Questions</h3><p><strong>Is copy trading on Solana profitable?</strong><br>It can be, but only if the wallet you copy makes more than the copy gives up. In our replay of 2,023 Pump.fun trades, a copy two slots behind did worse than the original 94% of the time. Copy wallets that are profitable without being first.</p><p><strong>How do I find a good wallet to copy?</strong><br>Start from coins that did well and look at their early holders who were not the creator. Look for names that show up again and again, then run the five checks: own trades, same wallet for buy and sell, real size, hold time a copy can match, and profit without the best trade.</p><p><strong>How many wallets should I copy?</strong><br>Start with one to three. More wallets mean more open positions at once and more overlap on the same coins. Add wallets one at a time once the first ones have a week of results.</p><p><strong>Should I copy the wallets of launch snipers?</strong><br>Usually not. In our replay, copies of wallets that bought in the first blocks went from 55% winners to 16%. Their edge is the entry price, and a copy cannot get it.</p><p><strong>What size should I use?</strong><br>Something you are fine losing while you test, like 0.05 to 0.1 SOL. Or a share of the wallet’s buy with a cap. Do not copy a wallet that buys 0.002 SOL at 0.1 SOL a time.</p><p><strong>Should my bot sell when the copied wallet sells?</strong><br>Yes, in most cases. Holding past the wallet’s sell made the average result much worse in our replay. Keep your own stop-loss and timer as a backup.</p><p><strong>What if the wallet buys and never sells?</strong><br>Then you need your own exits. It may be holding long, or it may be selling from a second wallet you cannot see. A timer and a stop-loss stop you from being stuck.</p><p><strong>Do I need to copy a wallet on every launchpad?</strong><br>No. Most wallets trade mainly on one or two venues. Tick the venues that wallet actually uses. If it only trades Pump.fun coins and PumpSwap pools, copying it on other venues adds nothing and only widens what your filters have to catch.</p><p><strong>How long should I test a wallet before trusting it?</strong><br>At least a week at a small size, with at least 20 finished trades. Then compare your copies to the wallet’s own trades on the same coins. If your results are close to theirs, minus a small gap, the setup works. If they are far apart, find out why before you size up.</p><p><strong>Can a copy bot be as fast as the wallet it follows?</strong><br>No. It has to see the trade first. The best a copy can do is land a slot or two later, every time.</p><h3>How we collected the data</h3><p>On November 19, 2026, we asked a Solana RPC node for the latest signatures on the Pump.fun program, 6EF8rrecthR5Dkzon8Nwu78hRvfCKubJ14M5uBEwF6P. The last 3,026 successful ones covered 10:56:44 to 10:57:45 UTC. We fetched 1,200 of them in full, every second one, read the program log to tell buys from sells, and counted the wallet that signed and paid for each. 95 wallets bought or sold. We kept the 15 that showed up most.</p><p>For each of the 15 we read up to 1,000 of its latest signatures to measure pace, wallet age and failure rate, and fetched its last 150 transactions, or all of them if it had fewer. For every transaction we took the wallet’s change in SOL, counting wrapped SOL as SOL, and its change in each token. A buy is a transaction where the wallet’s token balance went up and its SOL went down. A sell is the other way round. A round trip is a coin the wallet both bought and sold inside the window. Profit is the SOL change across that coin’s transactions, so network fees, priority fees and tips are included. Hold time runs from the first buy to the last sell.</p><p>Limits: one minute of Pump.fun activity picks out the busiest wallets, not the best ones. 150 transactions is a short window for the fast wallets, often under half an hour. Coins still held at the end of a window are left out of profit. We read only what is on chain, so we can describe what the wallets did, not who runs them or why. The replay figures come from our separate replay of 2,023 Pump.fun trades earlier in November 2026.</p><h3>Sources</h3><ul><li><a href="https://github.com/pump-fun/pump-public-docs">Pump.fun public docs on GitHub</a>, for the program, its instructions and the bonding curve.</li><li><a href="https://solana.com/docs/rpc/http/getsignaturesforaddress">Solana RPC: getSignaturesForAddress</a>, used to list each wallet’s history.</li><li><a href="https://solana.com/docs/rpc/http/gettransaction">Solana RPC: getTransaction</a>, used to read balance changes and logs.</li><li><a href="https://solana.com/docs/core/fees">Solana docs on fees</a>, for base fees and priority fees.</li><li><a href="https://solscan.io/account/9RDiYHEincnsgA6LYe3U7GcoHVtBjmERyp5dgpjdeLpx">Wallet A on Solscan</a>, the wallet that only sold.</li><li><a href="https://solscan.io/account/494pzeUAKvksnas1HANZS56Y84kCZzNmRvSfU9Kdjw1U">Wallet B on Solscan</a>, the wallet that only paid.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cdc5e833bfde" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Boop Sniper Bot: 7 Myths About Sniping Boop, Checked on Chain]]></title>
            <link>https://medium.com/@dexsnipecom/boop-sniper-bot-7-myths-about-sniping-boop-checked-on-chain-d15ade1af5d7?source=rss-2654b58e3d1e------2</link>
            <guid isPermaLink="false">https://medium.com/p/d15ade1af5d7</guid>
            <category><![CDATA[cryptocurrency]]></category>
            <category><![CDATA[solana-blockchain]]></category>
            <category><![CDATA[memecoins]]></category>
            <category><![CDATA[trading-bot]]></category>
            <category><![CDATA[crypto-trading]]></category>
            <dc:creator><![CDATA[DexSnipe]]></dc:creator>
            <pubDate>Tue, 17 Nov 2026 07:18:07 GMT</pubDate>
            <atom:updated>2026-11-17T07:18:07.797Z</atom:updated>
            <content:encoded><![CDATA[<h3>We re-read every Boop curve and staker vault on November 17, 2026.</h3><p>Most advice about sniping Boop was written in May 2025, and it still repeats the same fears. We checked seven of them against the chain this week. Most were wrong, and the real risk on Boop turned out to be somewhere else.</p><figure><a href="https://www.dexsnipe.com/"><img alt="DexSnipe banner reading Follow the creators, not the feed, with the address https://www.dexsnipe.com/ and an Open dexsnipe.com button. A dashboard illustration lists four snipers: Creator watch on Boop and Bags following a wallet with 2 graduates, Fresh launch on Boop, Pump.fun and LetsBonk with a 5% dev cap, Migration on Boop and Raydium LaunchLab with a dev-sell exit, and a Copy trade sniper on 2 wallets, switched off." src="https://cdn-images-1.medium.com/max/1024/1*d39MSbra54aPiGTt8VLXNg.png" /></a><figcaption>DexSnipe is our own Solana sniper bot service.</figcaption></figure><h3>The short answer</h3><ul><li><strong>Boop is almost silent.</strong> One new Boop coin in the last 30 days, and 202 successful transactions on the whole launchpad program. The last graduation was on September 22, 2025.</li><li><strong>Stakers did not dump graduates at the open.</strong> The staker vault opens about a minute after graduation, and only 18% of claims ever landed in the first hour.</li><li><strong>Almost nobody claimed.</strong> 89.5% of the full staker allocations are still sitting in their vaults today.</li><li><strong>Those tokens did not vanish.</strong> All 1,038 vaults passed their clawback date months ago. None has been clawed back.</li><li><strong>The real risk is the creator.</strong> On the only new Boop coin this month, the creator bought 5.6% of supply in the launch transaction and sold all of it within 92 seconds.</li></ul><h3>First, what is a Boop sniper bot?</h3><p>Boop is a Solana launchpad that opened in May 2025. Every Boop coin starts on Boop’s own bonding curve, a program written for the launchpad. Buyers pay SOL into the curve, and the price rises with every buy. When the curve holds 86 SOL, the coin graduates. A single transaction takes a 6 SOL graduation fee, moves the rest of the SOL and a block of tokens into a new Raydium CPMM pool, and sends 5% of the supply toward people who stake the BOOP token.</p><p>A Boop sniper bot watches for those events and buys on rules you set. There are four common ways to aim one:</p><ul><li><strong>Fresh launch.</strong> Buy in the first seconds of a new Boop curve that passes your filters.</li><li><strong>Migration.</strong> Buy when a curve graduates and the Raydium pool opens.</li><li><strong>Creator watch.</strong> Buy the next coin that a wallet you name launches.</li><li><strong>Copy trade.</strong> Buy when a wallet you follow buys a Boop coin.</li></ul><p>The part that made Boop different, and the part most sniper guides worried about, is the staker share. Every coin that graduated sent 50 million tokens, 5% of its supply, to a claim vault for BOOP stakers. Thousands of wallets could claim. The fear was simple: a sniper buys the open, and a wave of free tokens lands on top of it.</p><p>So we went and looked. The myths below are the claims we kept seeing in Boop guides, threads and bot docs. Each one is followed by what the chain shows.</p><h3>Myth 1: “Boop is one of the busiest launchpads on Solana”</h3><p><strong>What the chain shows: it was, for about a month. It is almost silent now.</strong></p><p>Boop’s launchpad program holds one account per bonding curve. On November 6, 2026 we counted 49,931 of them. On November 17 there were 49,932. One new Boop coin in eleven days, and it was the only one in the last 30 days.</p><p>Graduations tell the same story. Each graduation opens one staker distributor, so the distributors are a clean record of when Boop coins made it to a pool:</p><ul><li><strong>May 2025:</strong> 927 graduations.</li><li><strong>June 2025:</strong> 56.</li><li><strong>July 2025:</strong> 14.</li><li><strong>August 2025:</strong> 34.</li><li><strong>September 2025:</strong> 7, the last on September 22.</li><li><strong>October 2025 to today:</strong> none.</li></ul><figure><img alt="Bar chart of Boop graduations by month, 1,038 in total: May 2025 927, June 56, July 14, August 34, September 7, October 2025 to now 0. Last graduation September 22, 2025. A side panel for October 18 to November 17, 2026 shows 1 new coin on November 16, 2026, 202 successful transactions on the whole program, 67 different wallets, 21 of them in the last 7 days." src="https://cdn-images-1.medium.com/max/1024/1*SKGwz9kc765oeC1F8ob3ag.png" /></figure><p>We also pulled every transaction on the launchpad program for the 30 days from October 18 to November 17, 2026. There were 285 in total and 202 that succeeded. They came from 67 different wallets and fell on 25 different days. 116 of them included a buy on a Boop curve and 86 included a sell. Only 21 came in the last seven days.</p><p>A Fresh launch sniper pointed only at Boop would have fired once this month. A Migration sniper pointed only at Boop has not had anything to fire on for more than a year.</p><p>This does not mean the Boop program is broken. Curves still work, coins can still be launched, and someone launched one on November 16. It means the volume is gone. If you came looking for a Boop sniper because of how busy it looked in May 2025, the honest answer is that the crowd moved on.</p><h3>Myth 2: “Stakers dump every graduate the moment the pool opens”</h3><p><strong>What the chain shows: the staker tokens are not even claimable at the open, and most claims came much later.</strong></p><p>Here is how a Boop graduation actually ran, using TELE, a typical early graduate from May 10, 2025. Every step is a transaction on Solscan.</p><ol><li><strong>01:27:10 UTC.</strong> Graduation. 6 SOL fee taken. 80 SOL and 206.9 million TELE go into a new Raydium CPMM pool. 50 million TELE go to Boop’s distribution wallet, not to stakers.</li><li><strong>01:27:29.</strong> The pool’s LP tokens are locked, 19 seconds after graduation.</li><li><strong>01:28:18.</strong> The staker distributor is opened, with 3,006 wallets on its claim list.</li><li><strong>01:28:26.</strong> The 50 million TELE move into the claim vault.</li><li><strong>01:45:58.</strong> The first sizeable claim, about 191,000 TELE.</li></ol><p>So for the first minute after the pool opened, no staker could claim anything. The first real claim came about 19 minutes after the open. On a fresh pool, the first minute is where a migration sniper does its work. The staker wave the guides warned about did not exist in that minute.</p><p>Across a sample of 90 staker vaults spread evenly through Boop’s history, the pattern held:</p><ul><li><strong>Median first claim:</strong> about 2 minutes after the vault opened.</li><li><strong>Fastest tenth of first claims:</strong> under 40 seconds after the vault opened.</li><li><strong>Claims in the first hour:</strong> 18% of all claims.</li><li><strong>Claims in the first day:</strong> 67%.</li><li><strong>Claims in the first week:</strong> 93%.</li></ul><p>Some stakers were clearly running bots of their own and claimed within seconds. But the claims were spread over hours and days, not stacked on the open. A sniper holding for a few minutes saw a trickle at most.</p><h3>Myth 3: “The 5% airdrop gets claimed and sold”</h3><p><strong>What the chain shows: about one token in ten was ever claimed. The ones that were claimed were mostly let go.</strong></p><p>This is where the fear was furthest from the record. Boop’s distribution wallet opened 1,038 staker distributors. Each one records how many wallets were on its claim list and how many have claimed. On November 17, 2026 the totals were:</p><ul><li><strong>Places on claim lists:</strong> 2,617,287 across all 1,038 vaults.</li><li><strong>Claims made:</strong> 42,948. That is 1.6% of the places.</li><li><strong>Median vault:</strong> 2,644 wallets on the list, 13 claims.</li><li><strong>Full 50 million allocations still unclaimed:</strong> 89.5%, across 983 vaults.</li></ul><figure><img alt="Chart of the staker share on 983 Boop coins that sent the full 50 million tokens to a staker vault, read November 17, 2026: 89.5% still in the vaults and 10.5% claimed. 2,617,287 places on claim lists, 42,948 claims made, 1.6%. The median vault has 2,644 wallets listed and 13 claims. 1,038 of 1,038 vaults are past their one-year clawback date, 0 clawed back, about 44 billion tokens left, about 4.5% per coin." src="https://cdn-images-1.medium.com/max/1024/1*_tt5S2DZseSesfDJwDygbQ.png" /></figure><p>The claim count was exactly the same on November 6 and on November 17. Nobody claimed anything in between. For practical purposes, the staker airdrop on old Boop coins has stopped moving.</p><p>Why did so few claim? The chain does not say, but the shape of the numbers suggests it. A staker with a small BOOP position was on hundreds of claim lists, one per graduation. Each claim was a small slice of a coin they may never have heard of, and each one cost a transaction and a new token account. In May 2025 alone, 927 coins graduated. Most stakers simply did not bother.</p><p>The part of the myth that holds is what claimers did next. In our earlier sample of 108 early claims on 20 coins, 88 claiming wallets no longer held any of the coin when we checked. So the people who did claim mostly sold or moved the tokens. But the amounts were small. The median claim was about 219,000 tokens, around 0.02% of supply. Even the median coin’s total claims, about 2.8 million tokens, would move a fresh 80 SOL pool by only a couple of percent if sold all at once, and they were not sold all at once.</p><h3>Myth 4: “Unclaimed staker tokens are gone or burned”</h3><p><strong>What the chain shows: they are still in the vaults, and every vault can now be emptied in one transaction.</strong></p><p>Each Boop distributor records a clawback date. On all 1,038 opened by Boop’s wallet, that date is one year after the vault opened. After that date, the account allows everything still in the vault to be moved to a clawback receiver written into the account. For Boop’s vaults, that receiver is a token account owned by the same distribution wallet that received the 50 million at graduation.</p><p>On November 17, 2026:</p><ul><li><strong>Vaults past their clawback date:</strong> all 1,038.</li><li><strong>Vaults clawed back:</strong> 0.</li><li><strong>Tokens still in the 983 full vaults:</strong> about 44 billion, roughly 4.5% of each coin’s supply on average.</li></ul><p>So for a typical early Boop graduate, about 4.5% of the supply sits in one account that can move in a single transaction. We do not know what Boop plans to do with clawed-back tokens, and nothing on chain says they would be sold. But if you hold or plan to buy an old Boop coin, that block exists, and it is larger than most individual holders.</p><p>There is one exception. From July 6, 2025, Boop changed the airdrop. The next 55 graduations still sent 50 million tokens to the distribution wallet, but only 5 million went into the claim vault and the other 45 million were burned in the same transaction. On those coins, the most stakers could ever claim is 0.5% of supply. So “burned” is true for 45 million tokens on 55 coins. It is not true for the 983 coins before them.</p><p>How to check one coin yourself:</p><ol><li>Find the coin’s staker distributor on Solscan. It is an account owned by Boop’s distributor program, opened a minute or so after the graduation transaction.</li><li>Open its token vault and read the balance.</li><li>Compare it with 50 million, or 5 million for coins that graduated after July 6, 2025.</li></ol><p>That one number tells you how much of the staker share is still waiting.</p><h3>Myth 5: “Every Boop curve works the same way”</h3><p><strong>What the chain shows: the mechanics are the same, but the fee is not, and the fee mattered.</strong></p><p>Most of a Boop curve is uniform: a supply of one billion tokens, a graduation target of 86 SOL, a 6 SOL graduation fee, and about 30 SOL of virtual reserves that set the starting price. The trading fee is the part that varies. Every Boop curve charges either 1% or 2% a trade.</p><ul><li><strong>1% curves:</strong> 45,752 curves, 1,021 graduated. A graduation rate of 2.2%.</li><li><strong>2% curves:</strong> 4,179 curves, 29 graduated. A graduation rate of 0.7%.</li></ul><p>The 1% curves graduated three times as often. Part of that is cost: on a 2% curve, a buy and a sell cost 4% before network fees, which eats most of a quick move. Part is probably who picks the higher fee. A creator who raises the fee before a coin has found any buyers is aiming at the wrong thing.</p><p>The fee is written into the curve account from the start, so a sniper can see it before the first buy. On the record, skipping 2% curves costs you very few graduates.</p><p>A smaller myth hides here too: “every Boop mint ends in boop.” Of the 49,931 curves we read on November 6, 49,554 mints ended in those letters. Most do. Not all. The only new Boop coin this month has a mint that does not end in boop. The certain test is the curve account on Boop’s program, not the mint address.</p><h3>Myth 6: “On Boop, the fastest sniper wins”</h3><p><strong>What the chain shows: on the one new Boop coin this month, the fastest buyers lost, and the creator decided the outcome.</strong></p><p>Here is that coin, launched on November 16, 2026, read transaction by transaction. Times are seconds after the launch transaction at 09:07:36 UTC.</p><ul><li><strong>0 seconds.</strong> The creator launches the coin and buys in the same transaction: about 1.8 SOL for 56 million tokens, 5.6% of the supply.</li><li><strong>About 2 seconds.</strong> Two buyers land in the same block: one for about 0.205 SOL, one for about 0.0045 SOL.</li><li><strong>About 5 seconds.</strong> A tiny buy of about 0.003 SOL through a routing program.</li><li><strong>About 10 seconds.</strong> A buy of about 0.25 SOL.</li><li><strong>About 31 seconds.</strong> The creator sells part of the bag for about 0.45 SOL.</li><li><strong>About 92 seconds.</strong> The creator sells the rest for about 1.29 SOL. Their token account is now closed.</li><li><strong>About 27 minutes.</strong> The 0.205 SOL buyer sells for about 0.171 SOL.</li><li><strong>The next day.</strong> The 0.25 SOL buyer sells for about 0.197 SOL.</li></ul><figure><img alt="Timeline of the only new Boop coin this month, launched November 16, 2026 at 09:07:36 UTC. At 0 s the creator buys 5.6% of supply for 1.8 SOL. At 2 s two buyers pay 0.205 and 0.0045 SOL, at 10 s another pays 0.25 SOL. The creator sells for 0.45 SOL at 31 s and 1.29 SOL at 92 s. The 2-second buyer gets 0.171 SOL back, about -17%, the 10-second buyer 0.197 SOL, about -22%. A 5% dev cap would have skipped it." src="https://cdn-images-1.medium.com/max/1024/1*RNqEa_i5eHud8xQzWCdHIA.png" /></figure><p>The two buyers who put real SOL in and later sold both got back less than they paid: about 17% less for the 2-second buyer and about 22% less for the 10-second buyer. The creator lost a little, about 0.09 SOL after fees. The creator’s wallet was not new either. It had signed more than 1,000 transactions in the three days before this launch.</p><p>One coin is one coin, and we are not claiming every Boop launch went like this. But it shows the thing that decided the outcome. Speed did not matter here. What mattered was that the creator held 5.6% of the supply from the first block and sold it into the first buyers within a minute and a half. A sniper with a dev holding cap under 5% would have skipped this coin. A sniper with a dev-sell exit would have sold at the 31-second mark, when the creator’s first sell landed.</p><p>This is the same lesson we found on Pump.fun, with far more coins behind it, in <a href="https://medium.com/@dexsnipecom/pump-fun-sniper-bot-filters-we-checked-807-new-coins-to-see-which-rules-work-5c7a932f9e9e">Pump.fun sniper bot filters: we checked 807 new coins to see which rules work</a>. Dev filters did more than speed settings there too.</p><h3>Myth 7: “A Boop sniper needs special settings for the staker airdrop”</h3><p><strong>What the chain shows: the airdrop is too slow and too small to matter for a short trade. The creator and the holder spread matter more.</strong></p><p>Put the numbers together. The vault is not claimable for the first minute after graduation. The median first claim came about two minutes after that. Only about one claim in five landed in the first hour, and the median claim was 0.02% of supply. If your exits are measured in minutes, the staker share is noise next to ordinary trading in a pool that just raised 86 SOL.</p><p>There is one practical detail worth knowing. At graduation, the 50 million staker tokens land in a single distribution wallet before they move to the vault. For about a minute, that wallet is one of the largest holders of the coin, with 5% of supply. If your bot’s holder check counts every wallet, that wallet will show up. Set your top-10 holder cap with that in mind, or you may skip graduates for a reason that has nothing to do with the creator.</p><p>For coins that graduated before July 2025, the vault is the bigger holder detail. It can hold close to 5% of supply for good, and it can be clawed back in one transaction. Treat it like any other large wallet when you read the holder list.</p><h3>So what is the real risk on Boop?</h3><p>The same one as on every launchpad: who launched the coin, and how much of it they hold. Boop’s record makes that clear.</p><ul><li><strong>355 wallets made 44% of all Boop launches.</strong> Their coins graduated at 1.2%, against 2.6% for one-time creators and 3.1% for wallets with 2 to 9 launches.</li><li><strong>The busiest single wallet launched 2,662 Boop coins.</strong></li><li><strong>719 creator wallets graduated at least one Boop coin.</strong> 111 graduated two or more, and one graduated 54.</li></ul><p>So the bulk launchers filled the feed with coins that rarely went anywhere, and a small group of creators kept taking coins to 86 SOL. That second group is exactly what a Creator watch sniper is for. If one of those wallets launches again, on Boop or anywhere else you have ticked, you want to be in it. If a bulk launcher drops its 2,663rd coin, you do not.</p><p>We saw the same split on Believe, where one-time creators drew 26 times the trading of bulk launchers. The full numbers are in <a href="https://medium.com/@dexsnipecom/believe-sniper-bot-in-2026-12-questions-about-sniping-believe-answered-from-the-chain-7e6211300cc0">Believe sniper bot in 2026: 12 questions about sniping Believe, answered from the chain</a>.</p><h3>Which trigger fits Boop now?</h3><p>The four triggers behave very differently on a launchpad this quiet. Here is how we would think about each one.</p><p><strong>Fresh launch.</strong> On a busy launchpad, Fresh launch is the default. On Boop it fires about once a month. When it does fire, the coin has almost no other buyers, so your own buy moves the price a lot and there is nobody behind you to sell to. The November 16 coin had four outside buyers in its whole life. If you keep Fresh launch on Boop, keep the size small, keep the dev cap tight, and keep Boop as one venue among several, never the only one.</p><p><strong>Migration.</strong> A Boop Migration sniper buys the moment a curve reaches 86 SOL and the Raydium pool opens. That has not happened since September 22, 2025. On November 6, 2026, no live Boop curve held more than about 25 SOL. A Migration sniper on Boop alone is a sniper that waits. If you want migration snipes, point the same sniper at launchpads where curves still complete, and leave Boop ticked in case one does.</p><p><strong>Creator watch.</strong> This is the trigger that fits Boop’s record best. 719 wallets took at least one Boop coin to a pool, and 111 did it more than once. Some of those wallets may still be launching coins, on Boop or elsewhere. A Creator watch sniper follows the wallet, not the launchpad, so if one of them launches on any venue you have ticked, you buy it. Building that list takes an evening with an explorer: find graduated Boop coins, read who created them, and keep the wallets with more than one graduation.</p><p><strong>Copy trade.</strong> The same logic applies to buyers. A wallet that bought early and sold well on Boop in 2025 may be trading on other launchpads now. Copy trade follows that wallet to any venue you tick. On Boop itself it will almost never fire, because the wallets you would copy have left.</p><p>So the short version: on Boop today, follow people, not the launchpad. The launchpad has gone quiet. Its best creators and traders may not have.</p><h3>Three mistakes we see with Boop snipers</h3><ol><li><strong>Running a Boop-only sniper.</strong> It sits idle for weeks, and when it fires it buys a coin nobody else is buying. Tick Boop next to active venues instead.</li><li><strong>Skipping graduates because of the staker share.</strong> The staker tokens were never the threat they looked like. Skip a graduate because of its creator or its holders, not because 5% was set aside for stakers who mostly never claimed.</li><li><strong>Trusting old guides on volume.</strong> A guide from May 2025 describing hundreds of graduations a month was accurate then. Check the launchpad’s recent activity before you arm anything. It takes two minutes on Solscan: open the program, look at the newest transactions, and see how far back the last few go.</li></ol><h3>How to read a Boop coin in three minutes</h3><p>Before a buy, or before you add a creator to a watch list, check these in order. All of it is public and free.</p><ol><li><strong>Is it a Boop coin?</strong> Find its curve account on Boop’s launchpad program. The mint ending in boop is a hint, not proof.</li><li><strong>What is the curve fee?</strong> 1% or 2%. A 2% curve costs you 4% per round trip and graduated far less often.</li><li><strong>How much did the creator buy at launch?</strong> Look at the launch transaction. On the November 16 coin it was 5.6% of supply.</li><li><strong>What is the creator’s record?</strong> How many Boop coins has this wallet launched, and did any graduate? Hundreds with none is a bulk launcher.</li><li><strong>How full is the curve?</strong> A curve at 70 of 86 SOL is a different trade from one at 2.</li><li><strong>For graduates: what is left in the staker vault?</strong> Compare the vault balance with 50 million, or 5 million after July 6, 2025.</li><li><strong>For graduates: who else holds it?</strong> Ignore the Raydium pool and the staker vault, then look at what is left. A creator or a cluster of fresh wallets holding a large share matters more than any staker claim.</li></ol><figure><img alt="Checklist of seven checks to read a Boop coin in three minutes: 1 is it a Boop coin, check the curve account on Boop’s program; 2 the curve fee, 1% or 2%; 3 how much the creator bought, from the launch transaction; 4 the creator’s record of launches and graduations; 5 how full the curve is, SOL inside out of 86; 6 for graduates, what is left in the staker vault against 50 million or 5 million; 7 who else holds it, skipping the pool and vault." src="https://cdn-images-1.medium.com/max/1024/1*N50lazGXtKeKuTHjEyPfHA.png" /></figure><h3>Boop sniper settings, if you keep Boop ticked</h3><p>Boop is quiet, so we would not run a Boop-only sniper. Tick Boop next to launchpads that are still active, so a rare Boop launch from a good creator is not missed. These are starting points, not promises of any result.</p><ul><li><strong>Trigger:</strong> Creator watch on wallets that have graduated Boop coins before, or Fresh launch with Boop ticked next to busier venues.</li><li><strong>Buy size:</strong> 0.1 to 0.3 SOL. A thin curve moves a lot on each buy.</li><li><strong>Slippage cap:</strong> 10 to 15%.</li><li><strong>Priority fee and tip:</strong> modest. On a curve with almost no other buyers, there is little to race against.</li><li><strong>Dev holding cap:</strong> 5%. The November 16 creator held 5.6% from the first block.</li><li><strong>Skip bundled launches:</strong> on.</li><li><strong>Top 10 holder cap:</strong> 25 to 35%, leaving room for the distribution wallet at graduation.</li><li><strong>Max transfer tax:</strong> keep it low; it costs nothing on normal coins.</li><li><strong>Take-profit:</strong> two or three steps, for example +40% sell half and +120% sell the rest.</li><li><strong>Stop-loss:</strong> 20 to 25%.</li><li><strong>Dev-sell exit:</strong> on. This is the one that would have mattered on November 16.</li><li><strong>Timer:</strong> 15 to 30 minutes. With so few traders, a coin that has not moved in a quarter hour rarely does.</li></ul><p>Count the curve fee in every target. On a 1% curve a round trip costs about 2% before network fees; on a 2% curve about 4%. If you also want to catch graduates, remember that Boop coins move to Raydium CPMM pools, which work like the pools we described for Pump.fun migrations in <a href="https://medium.com/@dexsnipecom/pumpswap-sniper-bot-10-mistakes-that-turn-a-pump-fun-migration-snipe-into-a-loss-871f672d0819">PumpSwap sniper bot: 10 mistakes that turn a Pump.fun migration snipe into a loss</a>. Most of those mistakes apply to any migration snipe.</p><h3>Where DexSnipe fits</h3><p>DexSnipe is our own Solana sniper bot. Boop is one of its 12 venues, with the Fresh launch, Migration, Creator watch and Copy trade triggers, and you can tick Boop next to any other venue on the same sniper. Every sniper has the filters this guide leans on: a dev holding cap, a top-10 holder cap, a bundled-launch check and a transfer tax cap. Every sniper also has the exits: up to four take-profit steps, a stop-loss, a trailing stop, a timer and a dev-sell exit. They run on our server after you close the tab. Plans are flat: 1 SOL for 7 days, 3 SOL for 30 days or 5 SOL for 90 days, with no fee per trade. You still pay network fees, your own priority fee and tip, and the curve’s or pool’s trading fee.</p><p>If you want to compare Boop’s rules with the other launchpads side by side, we did that in <a href="https://medium.com/@dexsnipecom/solana-launchpads-compared-the-hidden-rules-of-letsbonk-jupiter-studio-bags-boop-believe-and-87cef7c5a86d">Solana launchpads compared: the hidden rules of LetsBonk, Jupiter Studio, Bags, Boop, Believe and more</a>.</p><h3>Questions</h3><p><strong>Is Boop still launching coins?</strong><br>Barely. One new coin in the 30 days to November 17, 2026, on November 16. The last graduation was on September 22, 2025.</p><p><strong>How much SOL does a Boop coin need to graduate?</strong><br>86 SOL in the curve, with a 6 SOL graduation fee. In the graduation we traced, 80 SOL went into the new Raydium CPMM pool.</p><p><strong>Do BOOP stakers dump graduated coins?</strong><br>Far less than people feared. 89.5% of the full staker allocations were never claimed, claims came over hours and days, and the median claim was 0.02% of supply.</p><p><strong>What happens to unclaimed staker tokens?</strong><br>They stay in the vault. Every vault’s clawback date has passed, and as of November 17, 2026 none had been clawed back.</p><p><strong>Did Boop change the staker airdrop?</strong><br>Yes. From July 6, 2025, only 5 million tokens went to the claim vault and 45 million were burned. 55 coins graduated under that rule.</p><p><strong>What fee do Boop curves charge?</strong><br>1% or 2% a trade. 1% curves graduated three times as often.</p><p><strong>How do I recognise a Boop coin?</strong><br>Most Boop mints end in boop, but not all. The certain test is a curve account on Boop’s launchpad program.</p><p><strong>Does DexSnipe take a fee on Boop trades?</strong><br>No. Plans are flat, with no fee per trade.</p><h3>How we collected the data</h3><p>On November 17, 2026, between about 07:00 and 07:30 UTC, we read Boop’s two programs through a standard Solana RPC.</p><ul><li><strong>Curves:</strong> we counted every account of Boop’s launchpad program (address starting boop8hVG) by type. 49,932 are bonding curves, one more than on November 6.</li><li><strong>Activity:</strong> we read every transaction on the launchpad program from October 18 to November 17, 2026. 285 in total, 202 successful. For each successful one we read the signer, the program steps it ran, and whether the signer’s SOL went up or down.</li><li><strong>The November 16 coin:</strong> we read its launch transaction and every transaction on its mint, then checked whether the creator still held a token account for it. They did not.</li><li><strong>Staker vaults:</strong> we read all 1,055 accounts of Boop’s distributor program (address starting boopEtkT), kept the 1,038 opened by Boop’s distribution wallet, and read each one’s claim list size, claim count, clawback date and clawback flag. We then read the balance of every vault. The layout matches the open-source Jito merkle distributor; we checked it against a vault we traced by hand.</li></ul><p>Monthly graduations, curve fees, creator groups, claim timing and the claimer sample come from our full Boop census of November 6, 2026, which read the same accounts.</p><p>What this data does not tell you:</p><ul><li><strong>Wallets are not people.</strong> One person can run many wallets, and one service can use one wallet for many people.</li><li><strong>“No longer held” is not “sold”.</strong> A claimer who moved tokens to another wallet looks the same as one who sold.</li><li><strong>Buys and sells are inferred.</strong> We read a transaction’s direction from the signer’s SOL change and the program steps. Routed trades can mix several steps.</li><li><strong>The November 16 coin is one coin.</strong> It shows how a creator can decide a launch; it does not show what every Boop launch does.</li><li><strong>We have not seen Boop’s distributor source code.</strong> Its accounts match Jito’s layout field for field, but its clawback rules could differ.</li></ul><h3>Sources and further reading</h3><ul><li>Decrypt, Crypto influencers lured by $350K airdrop to launch Solana meme coins (Boop’s launch and the staker share, May 2025): <a href="https://decrypt.co/317452/crypto-influencers-350k-airdrop-solana-meme-coins">decrypt.co</a></li><li>BingX Learn, What is Boop.fun (stakers receive 5% of each graduating coin’s supply): <a href="https://bingx.com/en/learn/article/what-is-boop-fun-and-how-to-claim-your-boop-token-airdrop">bingx.com</a></li><li>Boop, the launchpad: <a href="https://boop.fun/">boop.fun</a></li><li>Boop launchpad program on Solscan: <a href="https://solscan.io/account/boop8hVGQGqehUK2iVEMEnMrL5RbjywRzHKBmBE7ry4">solscan.io</a></li><li>Boop distributor program on Solscan: <a href="https://solscan.io/account/boopEtkTLx8x8moK7mMBQZUfzaEiA96Qn7gQeNdcQMg">solscan.io</a></li><li>Jito merkle distributor on GitHub (the account layout Boop’s vaults match): <a href="https://github.com/jito-foundation/distributor">github.com/jito-foundation</a></li><li>TELE graduation transaction, May 10, 2025, on Solscan: <a href="https://solscan.io/tx/3MWHVnCmSD2UQ5yzyxUbHd7JxQYx9zger6rBtYNuagE8Ar8EY1jDoUHjj4Zj4589atykKbc9gVxzca6TYMP4Wfy4">solscan.io</a></li><li>TELE staker distributor on Solscan (3,006 wallets on the list, 14 claims): <a href="https://solscan.io/account/BV1C8DT9awWgJhuw5aMAx95XWDhSEfLTM3ASqxTX2Lv5">solscan.io</a></li><li>The November 16, 2026 Boop launch transaction on Solscan: <a href="https://solscan.io/tx/32JCzE8xbjGLRBYF6vgjnA47kh6sGy1TcycYYRs1aBUNiSaMyoW3jq7CeVxh39hcAswtJT4iUMpTCWopEe9Gmggg">solscan.io</a></li><li>Raydium CPMM program on GitHub (the pools Boop coins graduate into): <a href="https://github.com/raydium-io/raydium-cp-swap">github.com/raydium-io</a></li></ul><p>Everything on chain was read in November 2026 and can change. Nothing here is financial advice. Meme coins are very risky, and most new tokens lose most of their value. A sniper bot follows your rules; it does not make new coins safe. Only trade with money you can afford to lose.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d15ade1af5d7" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Believe Sniper Bot in 2026: 12 Questions About Sniping Believe, Answered From the Chain]]></title>
            <link>https://medium.com/@dexsnipecom/believe-sniper-bot-in-2026-12-questions-about-sniping-believe-answered-from-the-chain-7e6211300cc0?source=rss-2654b58e3d1e------2</link>
            <guid isPermaLink="false">https://medium.com/p/7e6211300cc0</guid>
            <category><![CDATA[cryptocurrency]]></category>
            <category><![CDATA[memecoins]]></category>
            <category><![CDATA[crypto-trading]]></category>
            <category><![CDATA[trading-bot]]></category>
            <category><![CDATA[solana-blockchain]]></category>
            <dc:creator><![CDATA[DexSnipe]]></dc:creator>
            <pubDate>Mon, 16 Nov 2026 18:34:06 GMT</pubDate>
            <atom:updated>2026-11-16T18:34:06.107Z</atom:updated>
            <content:encoded><![CDATA[<h3>People still search for a Believe sniper bot every day. So we went back to the chain on November 16, 2026 and read every Believe pool again, plus every trade on a Believe curve in the last 30 days. Here are the questions a sniper should ask, and what the chain says.</h3><p>If you want the answer in one line: Believe has not launched a single coin since December 25, 2025. We counted every pool created under Believe’s Meteora configs on November 16, 2026, and the number was 31,354, exactly the same as ten days earlier and exactly the same as last Christmas. In the last 30 days, all of Believe’s curves together saw 53 successful trades from 19 wallets, and most of those were people selling what they had left. A Believe sniper bot will still fire if Believe ever starts launching again. Until then it has nothing to buy. That does not make Believe useless to a sniper. It is one of the few launchpads with a complete record, start to finish, and the questions below show what that record teaches about every launchpad that is still running.</p><figure><a href="https://www.dexsnipe.com/"><img alt="DexSnipe banner reading Snipe where coins still launch, with the address https://www.dexsnipe.com/ and an Open dexsnipe.com button. A dashboard illustration lists four snipers: Creator watch on Believe and Bags following one wallet with a 10% dev cap, Fresh launch on Bags, Jupiter Studio and Believe, a patient Fresh launch sniper that waits for a 20% dip with a low tip, and a Copy trade sniper on 3 wallets, switched off." src="https://cdn-images-1.medium.com/max/1024/1*Nj-Eoh9sk24DB0OeVOqD_A.png" /></a><figcaption>DexSnipe is our own Solana sniper bot service.</figcaption></figure><h3>The short version</h3><ul><li><strong>No new Believe coins since December 25, 2025.</strong> All 712 of Believe’s pool configs together hold 31,354 pools. The newest one opened at slot 389,069,828, on Christmas Day 2025.</li><li><strong>Almost no trading either.</strong> 53 successful transactions touched a Believe curve config between October 17 and November 16, 2026. Only 4 of them came in the last 7 days.</li><li><strong>Most unfinished curves are empty.</strong> 30,776 Believe coins never completed their curve. The median one holds no SOL at all. Only 82 hold 1 SOL or more.</li><li><strong>578 Believe coins did complete their curve.</strong> They moved to Meteora pools and trade there like any other pool, outside the launch window a sniper is built for.</li><li><strong>Being first on Believe was expensive by design.</strong> 93% of launches charged 50% on the opening trade, falling to 2% after about a minute.</li><li><strong>The lesson travels.</strong> Count money, not launches. Follow creators, not tickers. Read the fee schedule. And check that a launchpad is still alive before you arm a sniper on it.</li></ul><h3>1. What is a Believe sniper bot?</h3><p>Believe was a Solana launchpad where you made a coin by replying to a post on X. Each coin started on a Meteora Dynamic Bonding Curve, often called DBC. Buyers paid SOL into the curve, and the price rose with every buy. When the curve had taken in enough SOL, the coin completed its curve and moved to a Meteora pool.</p><p>A Believe sniper bot is a program that watches for those events and buys on rules you set. There are four common ways to point one at Believe:</p><ul><li><strong>Fresh launch.</strong> Buy in the first blocks of every new Believe curve that passes your filters.</li><li><strong>Migration.</strong> Buy when a curve completes and the Meteora pool opens.</li><li><strong>Creator watch.</strong> Buy the next coin that a wallet you name launches.</li><li><strong>Copy trade.</strong> Buy when a wallet you follow buys a Believe coin.</li></ul><p>After the buy, the bot sells on exit rules: take-profit steps, a stop-loss, a trailing stop, a timer, or a rule to sell when the creator sells. None of that is special to Believe. What was special was how Believe charged the first buyers, who launched the coins, and how quickly the whole thing went quiet. Those are the parts this guide covers.</p><h3>2. Is Believe still launching coins in 2026?</h3><p>No. This is the question most Believe sniper pages skip, because most of them were written in mid-2025, when Believe was one of the busiest launchpads on Solana. Here is how we checked it on November 16, 2026, and how you can check it yourself.</p><p>Every Believe coin is a pool on Meteora’s Dynamic Bonding Curve program. Every pool points to a config account. Every config names a fee claimer, the account that collects the launchpad’s share of fees. All of Believe’s configs name the same fee claimer, an account that starts with 9supzj. So the whole history of Believe can be read in two steps:</p><ol><li>Ask the program for every config account whose fee claimer is Believe’s. On November 16 that returned 712 configs.</li><li>Ask the program for every pool created with one of those 712 configs. That returned 31,354 pools.</li></ol><p>31,354 is the same count we got on November 6, 2026, when we first did a full census of Believe. Each pool also stores the slot it became active in. The highest one is slot 389,069,828, which the chain dates to 12:47 UTC on December 25, 2025. That was the last Believe launch. Nothing has been created under Believe’s configs since.</p><figure><img alt="Bar chart of Believe launches by month in 2025: May 1,217 (8,026 SOL in curve fees), June 7,274 (33,439 SOL), July 4,900 (11,299 SOL), August 8,378 (2,720 SOL), September 3,810 (2,066 SOL), October 4,336 (1,908 SOL), November 1,276 (163 SOL), December 163 (15 SOL). A side panel shows 0 new launches from January 1 to November 16, 2026, the last launch on December 25, 2025 at 12:47 UTC, slot 389,069,828, and 53 trades on all Believe curves in the last 30 days." src="https://cdn-images-1.medium.com/max/1024/1*Jp3QR_goPJ9bMihsNJaQDQ.png" /></figure><p>The rise and fall is easy to see by month. Believe started on May 13, 2025, with 1,217 launches in its first half month. June had 7,274. August had the most of any month, 8,378. Then the numbers dropped: 3,810 in September, 4,336 in October, 1,276 in November and 163 in December. From January 2026 to today, zero.</p><p>The money fell faster than the launch count. Believe’s curves collected about 33,400 SOL in trading fees in June 2025 and about 15 SOL in December. We come back to that gap in question 9, because it is the most useful thing Believe teaches.</p><p>One more detail. Believe’s fee claimer account did not go silent in December. Its most recent transactions, when we looked, were in late July 2026, a few days apart. So someone was still moving fees around long after the last launch. But no new pool has used Believe’s configs in almost eleven months, and for a sniper that is the number that counts.</p><h3>3. Is anyone still trading Believe coins?</h3><p>Barely. A swap on a Meteora bonding curve has to include the pool’s config account, so the transaction history of Believe’s 712 configs is a record of every trade on every Believe curve. We pulled that history for the 30 days from October 17 to November 16, 2026.</p><ul><li><strong>53 successful transactions</strong> touched any Believe config in those 30 days.</li><li><strong>Only 4 of the 712 configs</strong> saw any trade at all. The other 708 were silent for the whole month.</li><li><strong>19 different wallets</strong> signed those 53 transactions.</li><li><strong>41 of them took SOL out</strong> of the signer’s side, about 5.1 SOL in total. In plain words, those were sells.</li><li><strong>7 of them put SOL in</strong>, about 1.1 SOL in total. Those were buys.</li><li><strong>Only 4</strong> came in the last 7 days.</li></ul><p>The busiest burst in the whole month was one wallet on November 9 selling small amounts 17 times in about a minute, each for about 0.005 SOL. That is someone clearing out leftovers, not a market. Several other trades came through routing programs, which means a holder asked an aggregator for the best price and the aggregator found the old curve.</p><p>On Believe, 53 trades took a month. A Fresh launch sniper has nothing to buy, and a Copy trade sniper pointed only at Believe would almost never see a signal.</p><h3>4. If I hold a Believe coin, can I still sell it?</h3><p>It depends on which kind of Believe coin you hold.</p><p><strong>If the coin completed its curve</strong>, it moved to a Meteora pool. 578 Believe coins did this. Those pools are ordinary pools now. You can sell through any Solana wallet or aggregator, and the price depends on how much is left in that pool.</p><p><strong>If the coin never completed its curve</strong>, the curve is the only place to sell. You sell tokens back into the curve and get SOL out of it. And here is the catch: you can only get out SOL that is still in the curve. We read the SOL reserve of all 30,776 unfinished Believe curves on November 16, 2026.</p><figure><img alt="Chart of SOL left in the 30,776 Believe curves that never completed, read November 16, 2026, on a log scale: 29,566 curves hold under 0.01 SOL, 824 hold 0.01 to 0.1 SOL, 304 hold 0.1 to 1 SOL, 78 hold 1 to 10 SOL and 4 hold 10 SOL or more. The median unfinished curve holds 0 SOL. All unfinished curves together hold about 490 SOL. 578 Believe coins completed their curve and moved to Meteora pools." src="https://cdn-images-1.medium.com/max/1024/1*ukwANNB5yzRCTjCxyPNxTQ.png" /></figure><ul><li><strong>29,566 curves hold less than 0.01 SOL.</strong> For these, there is nothing meaningful to sell into. The median unfinished Believe curve holds zero.</li><li><strong>824 hold between 0.01 and 0.1 SOL.</strong></li><li><strong>304 hold between 0.1 and 1 SOL.</strong></li><li><strong>78 hold between 1 and 10 SOL.</strong></li><li><strong>4 hold 10 SOL or more.</strong> The largest holds about 47.5 SOL.</li></ul><p>All together, about 490 SOL sits in the 30,776 unfinished curves. That is the whole exit for every holder of every Believe coin that never made it to a pool. If you hold one of the coins in the top few hundred, a sell still returns something. If you hold one of the other 29,566, the curve has already been emptied by the people who sold before you.</p><p>This is worth keeping in mind on any launchpad. On a bonding curve, early sellers take SOL out and late sellers find less of it. The exit is not a market that comes back later. It is a pot of SOL that only gets smaller once the buyers stop.</p><h3>5. What happened to Believe and LAUNCHCOIN?</h3><p>We can only tell you what is public and what is on chain. We do not know the team’s plans, and nothing here is a guess about the future.</p><p>What is public: in October 2025, Believe’s X account announced that LAUNCHCOIN, the token linked to the platform, was being upgraded to a new token called BELIEVE, with the upgrade window open until October 29, 2025. That was the last big announcement around the time launches fell off. November 2025 saw 1,276 launches, December 163, and then none.</p><p>What is on chain: the configs stopped getting new pools on December 25, 2025. The fee claimer stayed active into late July 2026. Curves that never completed still exist and can still be traded, but almost nobody does.</p><p>What it means for you: a sniper that was armed on Believe in 2025 and never checked again would have spent eleven months doing nothing. Nothing on a bot’s screen tells you a launchpad has stopped. The bot just waits. That is the reason for question 12.</p><h3>6. How did a Believe launch actually work?</h3><p>Believe’s idea, as Bankless described it in May 2025, was to turn posts on X into coins. You replied to Believe’s launch account with a name and a ticker, and a coin was created on the spot. A scout could also reply to someone else’s post and launch a coin for that idea. The founder behind the idea could claim half of the coin’s trading fees, even if somebody else had launched it.</p><p>On chain, every launch was a Meteora DBC pool created with one of Believe’s configs. Believe also gave its coins a recognisable mint address. Of the 31,354 Believe mints, 28,800 end in the letters BLV. If you see a mint ending in BLV on an explorer, it is very likely a Believe coin, though the certain test is the config and its fee claimer.</p><p>Each config also sets how much SOL the curve needs to complete. On Believe those targets were not all the same. The median pool needed about 203 SOL, and two thirds needed 200 SOL or more. That is a high bar compared with launchpads that complete at 85 SOL. It is one reason so few Believe coins, 578 out of 31,354, ever reached a pool.</p><p>If you want to compare Believe’s rules with the launchpads that are still running, we laid them side by side in <a href="https://medium.com/@dexsnipecom/solana-launchpads-compared-the-hidden-rules-of-letsbonk-jupiter-studio-bags-boop-believe-and-87cef7c5a86d">Solana launchpads compared: the hidden rules of LetsBonk, Jupiter Studio, Bags, Boop, Believe and more</a>.</p><h3>7. Why was being first on Believe a bad deal?</h3><p>Because Believe’s curves were built to charge the first buyers the most. Meteora’s bonding curve lets a launchpad set a fee schedule: a high fee at opening that falls over time. Believe used several, but one covered 29,025 of its 31,354 pools, 93%. It opened at 50% and fell by about 2.1% of the current fee every slot, for 150 slots.</p><p>In 2025 a Solana slot took about 0.4 seconds, so the schedule worked out like this:</p><ul><li><strong>At opening:</strong> 50% fee on the trade.</li><li><strong>About 4 seconds in:</strong> about 40%.</li><li><strong>About 12 seconds in:</strong> about 26%.</li><li><strong>About 20 seconds in:</strong> about 17%.</li><li><strong>About 30 seconds in:</strong> about 10%.</li><li><strong>About 40 seconds in:</strong> about 6%.</li><li><strong>After about a minute:</strong> 2%, where it stays.</li></ul><figure><img alt="Chart of Believe’s main launch fee, used by 29,025 of 31,354 pools, by time after opening, with the price rise needed to break even when selling at 2%: 50% at 0 seconds needs +104%, 40% at about 4 seconds needs +71%, 26% at about 12 seconds needs +38%, 17% at about 20 seconds needs +23%, 10% at about 30 seconds needs +13%, 6% at about 40 seconds needs +8%, and 2% after about 60 seconds needs +4%. Worked examples, not trade results." src="https://cdn-images-1.medium.com/max/1024/1*TqLoqofyeFuTmXZhCa9ZTQ.png" /></figure><p>Now do the math a sniper has to do. Say you buy at the opening fee of 50% and sell later at the normal 2% fee. Half of your buy goes to fees. To get your SOL back, the price has to roughly double before you sell. Buy at 26%, about 12 seconds in, and the price has to rise about 38% just to break even. Buy at 10% and you need about 13%. Buy after a minute and you need about 4%. These are rough figures from the fee formula. They leave out network fees and price impact, which only make it worse.</p><p>So the fastest buyer on Believe was not the winner. The fastest buyer paid for everybody else’s protection. The right move on a schedule like this was to wait, not to race.</p><p>Smaller groups of Believe pools used other schedules. 1,780 opened at 99% and fell to a few percent within about six seconds. 481 opened at 90%. A handful charged a flat 2% from the start. The habit that matters is the same for all of them: read the pool’s config before you decide how fast to be. We went through Meteora’s fee clocks second by second in <a href="https://medium.com/@dexsnipecom/meteora-sniper-bot-when-to-buy-a-launch-second-by-second-without-paying-the-anti-sniper-fee-5d0fea77d994">Meteora sniper bot: when to buy a launch, second by second, without paying the anti-sniper fee</a>.</p><h3>8. Who launched Believe coins, and who was worth following?</h3><p>17,908 different wallets launched at least one Believe coin. Most of them launched exactly one. But a small group launched far more, and the difference between the two groups is the clearest result in the whole record.</p><ul><li><strong>Wallets that launched one coin:</strong> 15,229 wallets, 15,229 coins. Average curve fees per coin: 3.39 SOL. 509 of these coins completed their curve.</li><li><strong>Wallets that launched 2 to 9 coins:</strong> 2,320 wallets, 7,504 coins. Average curve fees per coin: 0.92 SOL.</li><li><strong>Wallets that launched 10 or more:</strong> 359 wallets, 8,621 coins. Average curve fees per coin: 0.13 SOL. Only 13 of these coins completed their curve.</li></ul><p>A coin from a one-time creator drew, on average, 26 times the trading of a coin from a bulk launcher. The six busiest wallets launched 775, 560, 444, 432, 141 and 126 coins. Not one of their 2,478 coins completed its curve. A Fresh launch sniper with no creator filter would have bought every single one.</p><p>The coins that drew the most trading looked like small products, not jokes. The top five by curve fees were called Text Behind Image, hedgemony, JustWatched, Drilled and NeuroCheck. Each collected between about 184 and 353 SOL in curve fees and completed its curve. That fits what Believe set out to be: a place where builders launched coins for things they were making.</p><p>For a sniper the takeaway is simple. The useful unit was the creator, not the coin. Creator watch on a short list of wallets with a real record beat a fire-at-everything sniper by a wide margin. The same is true on Bags today, where every coin even names who earns its fees. We showed how to build that list step by step in <a href="https://medium.com/@dexsnipecom/bags-sniper-bot-how-to-build-a-watchlist-from-the-fee-record-every-bags-coin-leaves-behind-421549cc7cbe">Bags sniper bot: how to build a watchlist from the fee record every Bags coin leaves behind</a>.</p><h3>9. Why did Believe look busiest when it was already fading?</h3><p>Because launch counts and money told two different stories, and most people only looked at launch counts.</p><p>August 2025 had more launches than any other month on Believe: 8,378. A sniper set to fire on every new coin would have been busier than ever. But the curves collected about 2,700 SOL in fees that month, a twelfth of June’s 33,400. And 57% of August’s launches came from wallets that launched ten coins or more, against 1% in June. Six wallets alone made 1,066 of August’s launches.</p><p>In other words, the feed got fuller because a few wallets started launching in bulk, not because more people came to trade. The buyers had already started to leave. Fees per launch tell the story clearly: about 4.6 SOL per launch in June, about 0.44 SOL per launch in October.</p><p>Completed curves told the truth too. 292 coins launched in June went on to complete their curve. In August, with more launches than June, only 26 did. After November 9, 2025, no Believe launch completed its curve at all.</p><p>This is the single most useful habit Believe can teach. When you judge a launchpad, look at money per launch, the number of different creators, and how many curves complete. A rising launch count with falling fees and fewer creators is a warning, not an invitation.</p><h3>10. Should I keep a Believe sniper armed?</h3><p>It depends on how your bot charges and how you set it up.</p><p>If your bot takes a cut of every trade, an idle Believe sniper costs nothing, because it makes no trades. If your bot charges a flat plan, the question is whether Believe takes the place of a venue that is still launching. Many sniper bots, ours included, let one sniper watch several venues at once. So there is little reason to run a Believe-only sniper, and little cost to leaving Believe ticked next to venues that are still active.</p><p>Here is how we would handle it today:</p><ol><li><strong>Do not run a Fresh launch sniper on Believe alone.</strong> It will sit idle. If you want to catch a restart, tick Believe as one of several venues on a sniper you already run.</li><li><strong>Keep Creator watch for Believe creators with a record.</strong> The one-time creators whose coins drew real trading did not disappear. If one of them launches again on a venue you have ticked, a Creator watch sniper buys it.</li><li><strong>Keep a Migration sniper only if it covers other venues too.</strong> No Believe curve has completed since November 2025.</li><li><strong>Use Copy trade on wallets, not on platforms.</strong> A good wallet you followed on Believe is probably trading somewhere else now. Copy trade follows the wallet to any venue you tick.</li></ol><h3>11. Where does a Believe-style sniper go now?</h3><p>To launchpads that work the same way and are still launching. Three things made Believe what it was: a Meteora bonding curve, a launch fee that falls over time, and fees paid to named people. Other launchpads still have some or all of these.</p><ul><li><strong>Bags</strong> is the closest relative. It launches on the same Meteora curve and pays trading fees to named earners, including people who did not launch the coin. Its fee modes are different from Believe’s, so read the coin’s mode before setting targets.</li><li><strong>Jupiter Studio</strong> also launches on Meteora’s curve, and some of its configs open with a very high fee that falls over time, like Believe’s.</li><li><strong>Other Meteora DBC launchpads</strong> share the program, so the same checks work: read the config, read the fee schedule, read the creator’s history.</li></ul><p>Before you move a sniper to any of them, run the check in the next question. A launchpad that is busy today can look like Believe in August 2025: many launches, few buyers.</p><h3>12. How do I check that a launchpad is alive before I arm a sniper?</h3><p>Believe’s decline was visible on chain months before the launches stopped. Five checks would have caught it, and they take a few minutes on any launchpad with a free explorer.</p><figure><img alt="Checklist of five checks before arming a launchpad sniper, each with Believe’s answer: 1 is it still launching, last Believe launch December 25, 2025; 2 is money still moving, fees per launch 4.6 SOL in June 2025 and 0.44 SOL in October; 3 who is launching, 57% of August 2025 launches from wallets with 10 or more launches; 4 what the first buyer pays, 50% at opening and 2% after about a minute; 5 are coins still completing, 292 from June, 26 from August, none after November 9." src="https://cdn-images-1.medium.com/max/1024/1*YxS1JXb81vX5xoDtdpdxmQ.png" /></figure><ol><li><strong>Is it still launching?</strong> Open the launchpad’s program or config on Solscan and look at the newest pools or mints. On Believe today, the newest pool is from December 25, 2025. If you see no launch in the last day, do not arm a sniper on that venue alone.</li><li><strong>Is money still moving?</strong> Look at volume or fees per launch, not the number of launches. On Believe, fees per launch fell from about 4.6 SOL in June 2025 to about 0.44 SOL in October, while launches stayed above 4,000 a month.</li><li><strong>Who is launching?</strong> Look at the creators of the last 50 launches. If a few wallets made most of them, the feed is mostly bulk launches. On Believe in August 2025, 57% of launches came from wallets with ten or more launches.</li><li><strong>What does the first buyer pay?</strong> Read the fee schedule in the pool’s config. A 50% opening fee changes every speed and slippage setting you have.</li><li><strong>Are coins still completing?</strong> Count how many recent launches completed their curve. On Believe, only 26 August launches completed, against 292 in June. A launchpad where none complete for weeks has few buyers left.</li></ol><p>None of these needs a paid tool. They need an explorer and the habit of looking once a week. A sniper on a launchpad that has quietly emptied does not lose money by itself. It just stops earning, and it gives you the false feeling that you are in the market.</p><h3>Believe sniper settings, if you keep one armed</h3><p>These are starting points for a sniper that has Believe ticked in case it returns, or for a sniper on a launchpad with the same 50% fee schedule. They are not promises of any result.</p><ul><li><strong>Trigger:</strong> Creator watch on wallets with a real record, or Fresh launch with Believe ticked next to active venues.</li><li><strong>Buy size:</strong> 0.2 to 0.5 SOL for Creator watch, 0.05 to 0.1 SOL for Fresh launch.</li><li><strong>Slippage cap:</strong> 15% for Creator watch, 10% for Fresh launch.</li><li><strong>Priority fee and tip:</strong> low, for example 0.0005 SOL priority and 0.001 SOL tip. Paying extra to land in the most expensive slot makes no sense under a 50% opening fee.</li><li><strong>Buy the dip:</strong> worth turning on. Waiting for the price to fall back a set percentage also gives the launch fee time to fall.</li><li><strong>Dev holding cap:</strong> 5 to 10%.</li><li><strong>Top 10 holder cap:</strong> 25 to 35%.</li><li><strong>Skip bundled launches:</strong> on.</li><li><strong>Take-profit:</strong> two steps, for example +50% sell half and +150% sell the rest for Creator watch, lower for Fresh launch.</li><li><strong>Stop-loss:</strong> 25 to 30%.</li><li><strong>Timer:</strong> 15 to 60 minutes. Most late Believe launches saw almost no trading, and a timer frees the SOL.</li><li><strong>Dev-sell exit:</strong> on.</li></ul><p>Count the fee in every target. After the launch window, Believe curves charged 2% a trade, so a buy and a sell together cost about 4% before network costs. A take-profit at +5% on a Believe-style curve keeps very little. For more on which filters actually change results, see <a href="https://medium.com/@dexsnipecom/pump-fun-sniper-bot-filters-we-checked-807-new-coins-to-see-which-rules-work-5c7a932f9e9e">Pump.fun sniper bot filters: we checked 807 new coins to see which rules work</a>. The launchpad is different, but the filter logic is the same.</p><h3>Four lessons from a launchpad that stopped</h3><p>A launchpad that has stopped still gives you something no running launchpad can: the whole record, first launch to last, with nothing cut off and no survivors picked out. Four lessons stand out.</p><p><strong>Measure money, not launches.</strong> Believe’s launch count stayed above 3,800 a month until October 2025, long after its fees had collapsed. If you judge a venue by how many coins appear, you keep sniping it after the buyers have gone.</p><p><strong>Count creators, not coins.</strong> When a few wallets start launching hundreds of coins, the feed fills with tickets somebody else is buying. A creator filter, or a creator list built from wallets with a real record, is worth more than any speed setting.</p><p><strong>Read the fee schedule.</strong> On Meteora curves the first buyer can pay 50%, 90% or 99%. Speed is only an edge where the fee lets it be one.</p><p><strong>Watch the platform, not only the coin.</strong> Launchpads do not last forever. Spread snipers across venues, and check now and then that each one is still producing launches.</p><h3>More quick questions</h3><p><strong>How many coins were launched on Believe?</strong><br>31,354 under Believe’s Meteora configs, by 17,908 different creator wallets, between May 13 and December 25, 2025.</p><p><strong>How do I recognise a Believe coin?</strong><br>Most Believe mints end in BLV: 28,800 of 31,354. The certain test is the pool’s config, which names Believe’s fee claimer.</p><p><strong>How many Believe coins completed their curve?</strong><br>578. Their pools record that they migrated. The other 30,776 never completed.</p><p><strong>Can a sniper bot buy a Believe coin that already moved to a pool?</strong><br>Yes, like any token in any pool. But that is ordinary trading, not sniping. The launch window is long gone.</p><p><strong>Will Believe come back?</strong><br>We do not know, and we will not guess. If launches resume under Believe’s configs, a sniper with Believe ticked will see them. The check in question 12 will also show it within a day.</p><p><strong>Is there a free Believe sniper bot?</strong><br>If you find an open-source sniper bot on GitHub that mentions Believe, read its code before you give it a key, and check whether it was updated after Believe stopped launching. The checks in <a href="https://medium.com/@dexsnipecom/raydium-sniper-bot-on-github-9-checks-before-you-run-one-2566b90171e1">Raydium sniper bot on GitHub: 9 checks before you run one</a> apply to any venue.</p><p><strong>Does DexSnipe still list Believe?</strong><br>Yes. Believe is one of 12 venues, with the Fresh launch, Migration, Creator watch and Copy trade triggers. Plans are flat: 1 SOL for 7 days, 3 SOL for 30 days or 5 SOL for 90 days, with no fee per trade. You still pay network fees, your own priority fee and tip, and the coin’s trading fee.</p><h3>How we collected the data</h3><p>On November 16, 2026, around 18:00 to 18:30 UTC, we queried the Meteora Dynamic Bonding Curve program (address starting dbcij3) through a standard Solana RPC.</p><ul><li><strong>Configs:</strong> we listed every config account of the program and kept those whose fee claimer is Believe’s (address starting 9supzj). That gave 712 configs.</li><li><strong>Pools:</strong> for each config we listed every pool created with it and read each pool’s stored fields: creator, SOL reserve, activation slot and migration flag. That gave 31,354 pools, 578 marked as migrated.</li><li><strong>Last launch:</strong> the highest activation slot, 389,069,828, belongs to a pool created at 12:47 UTC on December 25, 2025. We confirmed the date from the creating transaction.</li><li><strong>Trading in the last 30 days:</strong> for each config we read its transaction history back to October 17, 2026. Every swap on a curve includes its config, so this covers every curve trade. We kept successful transactions and read each one: who signed it and whether the signer’s SOL went up or down.</li><li><strong>SOL left in curves:</strong> the SOL reserve field of each unfinished pool. We checked it against one curve’s SOL vault, which also holds uncollected fees, so the vault reads a little higher than the reserve.</li></ul><p>Monthly launch counts, creator groups, fee schedules and the top coins come from our full Believe census of November 6, 2026, which read the same configs and pools. Fees per month are the pools’ own running fee totals.</p><p>What this data does not tell you:</p><ul><li><strong>Wallets are not people.</strong> One person can use many wallets, and one service can use one wallet for many people.</li><li><strong>Fees are not profit.</strong> A fee total measures trading. It does not show who made or lost money.</li><li><strong>Sells and buys are inferred.</strong> We counted a transaction as a sell when the signer’s SOL went up. Routed trades can mix several steps.</li><li><strong>Times are approximate.</strong> The fee schedule is in slots. Seconds assume about 0.4 seconds per slot, Solana’s pace in 2025.</li><li><strong>The break-even figures are worked examples</strong> of the fee formula, not trade results.</li></ul><h3>Sources and further reading</h3><ul><li>Bankless, Believe: the Solana token launcher (launching by reply on X, founders claiming half of trading fees, May 2025): <a href="https://www.bankless.com/read/believe-solana-token-launcher">bankless.com</a></li><li>Believe on X, the LAUNCHCOIN to BELIEVE upgrade announcement, October 2025: <a href="https://x.com/believeapp/status/1978518886166245463">x.com/believeapp</a></li><li>CoinGecko Learn, What is Believe: <a href="https://www.coingecko.com/learn/what-is-believe-token-launchpad">coingecko.com</a></li><li>Bitget Academy, creating crypto by replying on X with Believe: <a href="https://web3.bitget.com/en/academy/what-is-launchcoin-and-how-to-create-crypto-by-replying-on-x-with-believe-app">web3.bitget.com</a></li><li>Meteora docs, What is DBC: <a href="https://docs.meteora.ag/core-products/dbc/what-is-dbc">docs.meteora.ag</a></li><li>Meteora docs, DBC fee scheduler (cliff fee, periods, reduction factor and the formula): <a href="https://docs.meteora.ag/core-products/dbc/fees/fee-scheduler">docs.meteora.ag</a></li><li>Meteora Dynamic Bonding Curve source code on GitHub (the pool and config layouts we read): <a href="https://github.com/MeteoraAg/dynamic-bonding-curve">github.com/MeteoraAg</a></li><li>Meteora Dynamic Bonding Curve program on Solscan: <a href="https://solscan.io/account/dbcij3LWUppWqq96dh6gJWwBifmcGfLSB5D4DuSMaqN">solscan.io</a></li><li>Believe’s fee claimer on Solscan: <a href="https://solscan.io/account/9supzj2Nbnjut3r9UFfWE2RiuAenQMoYaNtZXmmtm5oC">solscan.io</a></li><li>The last Believe pool, created December 25, 2025, on Solscan: <a href="https://solscan.io/tx/3Fr2gWzHzsWLyxVqBAD5isX6eLqi7xsoH6aLUtqBPcKhPohvhotrijptu7jvGLjQokaU5ALeWBtTUrH9QnA1i1UZ">solscan.io</a></li><li>The unfinished Believe curve with the most SOL left, on Solscan: <a href="https://solscan.io/account/9U7t5Pi5PyRA5fkyqZrCx4wK6PPJeHJcDbqqNz728Z2B">solscan.io</a></li></ul><p>Everything on chain was read on November 16, 2026, and can change if Believe’s configs are used again. Nothing here is financial advice. Meme coins are very risky, and most new tokens lose most of their value. A sniper bot follows your rules; it does not make new coins safe. Only trade with money you can afford to lose.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7e6211300cc0" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Bags Sniper Bot: How to Build a Watchlist From the Fee Record Every Bags Coin Leaves Behind]]></title>
            <link>https://medium.com/@dexsnipecom/bags-sniper-bot-how-to-build-a-watchlist-from-the-fee-record-every-bags-coin-leaves-behind-421549cc7cbe?source=rss-2654b58e3d1e------2</link>
            <guid isPermaLink="false">https://medium.com/p/421549cc7cbe</guid>
            <category><![CDATA[crypto-trading]]></category>
            <category><![CDATA[solana-blockchain]]></category>
            <category><![CDATA[trading-bot]]></category>
            <category><![CDATA[memecoins]]></category>
            <category><![CDATA[cryptocurrency]]></category>
            <dc:creator><![CDATA[DexSnipe]]></dc:creator>
            <pubDate>Mon, 16 Nov 2026 06:52:53 GMT</pubDate>
            <atom:updated>2026-11-16T06:52:53.455Z</atom:updated>
            <content:encoded><![CDATA[<h3>Every coin on Bags names the people who earn its trading fees, and the chain shows whether they ever collect. Here is a step-by-step way to turn that public record into a short list of launchers worth sniping, with the settings, the fee math and one real coin read from start to finish.</h3><p>Most launchpads give a sniper very little to go on in the first second of a launch: a name, a picture, a ticker and a bonding curve that just opened. Bags gives you one more thing. Every Bags coin carries a fee-share record that says who is supposed to earn from it, how much each of them gets, and how much they have actually collected so far. That record is public, it builds up over weeks, and most people who snipe Bags launches never read it. This guide shows how to read it, how to turn it into a watchlist of launcher wallets in seven steps, and how to set up a Bags sniper bot around that list instead of firing at every new coin. On the way we read one coin in full: a Bags coin that collected about 1,115 SOL in fees, named a single earner, and still had 430.36 SOL waiting for that person when we checked on November 16, 2026. Nobody had ever picked it up.</p><figure><a href="https://www.dexsnipe.com/"><img alt="DexSnipe banner reading Snipe the Bags launchers you checked, with the address https://www.dexsnipe.com/ and an Open dexsnipe.com button. A dashboard illustration lists four snipers: two Creator watch snipers on Bags, each following one launcher wallet with a 10% dev holding cap, a strict Fresh launch sniper on Bags with top 10 holders capped at 25%, and a Copy trade sniper on one wallet, switched off." src="https://cdn-images-1.medium.com/max/1024/1*_nGr2s7ZTqlJesSVoU5Y3w.png" /></a><figcaption>DexSnipe is our own Solana sniper bot service.</figcaption></figure><h3>The short version</h3><ul><li><strong>Bags coins pay their fees to named people.</strong> A launcher can name up to 100 fee earners. An earner can be a wallet, or a social account such as an X or GitHub handle that Bags turns into a wallet.</li><li><strong>Naming someone costs nothing, and they do not have to agree.</strong> That is why so many Bags coins are tributes: coins named after a person, launched by someone else.</li><li><strong>Whether the earner ever collects is public.</strong> When we read every fee account of the Bags Fee Share V2 program on November 6, 2026, 41% of the coins that had earned 1 SOL or more had never seen a single claim by anyone named on them.</li><li><strong>The launcher is paid either way.</strong> On new SOL configs, the wallet that pays for the launch gets its own 25% cut of the earners’ side of the fees, whether or not the named person ever shows up.</li><li><strong>So the useful unit is the launcher, not the coin.</strong> Read the record of last week’s coins, find the launchers whose coins kept trading and whose earners kept claiming, and point a Creator watch sniper at each one.</li><li><strong>Know the fee mode before you set targets.</strong> Bags has seven SOL fee modes. On the default one a buy and a sell on the curve cost about 4% together. On the most expensive one, about 19%.</li></ul><h3>What is a Bags sniper bot?</h3><p>Bags is a Solana launchpad. A new Bags coin starts on a Meteora Dynamic Bonding Curve. Tokens are sold for SOL along the curve, and the price rises with each buy. When the curve has raised enough SOL, 85 SOL in most fee modes, the coin graduates and moves to a Meteora DAMM v2 pool, where it trades like any other pool on Solana.</p><p>A Bags sniper bot watches for those events and buys on rules you set. There are four common ways to aim it:</p><ul><li><strong>Fresh launch.</strong> Buy in the first blocks of every new Bags curve that passes your filters.</li><li><strong>Migration.</strong> Buy when a curve completes and the DAMM v2 pool opens.</li><li><strong>Creator watch.</strong> Buy the next coin that a wallet you name launches.</li><li><strong>Copy trade.</strong> Buy when a wallet you follow buys a Bags coin, and sell when it sells.</li></ul><p>The curve itself works much like other Meteora launches. If you want the second-by-second side of that, we covered it in <a href="https://medium.com/@dexsnipecom/meteora-sniper-bot-when-to-buy-a-launch-second-by-second-without-paying-the-anti-sniper-fee-5d0fea77d994">our Meteora sniper bot guide</a>. And if you are still deciding which launchpad to snipe at all, <a href="https://medium.com/@dexsnipecom/solana-launchpads-compared-the-hidden-rules-of-letsbonk-jupiter-studio-bags-boop-believe-and-87cef7c5a86d">our launchpad comparison</a> lays out how Bags differs from LetsBonk, Jupiter Studio, Boop and Believe.</p><p>What makes Bags different is not the curve. It is who the coin pays, and the record that leaves. This guide is about using that record to decide which of those four triggers to use, and on whom.</p><h3>How does Bags fee sharing work?</h3><p>Every trade on a Bags coin pays a fee. Part of it goes to Bags, and part of it goes to the people named on the coin. Here is how the pieces fit, taken from Bags’ own documentation and its published program interface.</p><h3>The fee itself</h3><p>The launcher picks a fee mode when the coin is created. If they pick nothing, the coin uses the Default mode: a flat 2% on every trade, on the curve and after graduation. Before graduation that 2% is split evenly between a platform pool and a claimers pool. After graduation a quarter of it is added back into the pool’s liquidity, and the rest is split evenly between the same two pools.</p><h3>The claimers</h3><p>The claimers pool is shared by the earners in the coin’s fee-share config. A config can name up to 100 claimers, each with a share in basis points, and the shares add up to 100%. A claimer can be a Solana wallet. It can also be a social account, such as an X, Instagram or GitHub handle. Bags resolves the handle to an embedded wallet that the owner of that account can claim from by signing in.</p><p>That last part is the reason this guide exists. To name someone as the earner on a coin, a launcher needs only their handle. The named person does not need a Solana wallet, does not need to have heard of Bags, and does not need to agree.</p><h3>The deployer</h3><p>On SOL configs created with the current version of the Bags fee-share program, the wallet that pays to create the config is registered as the deployer. The deployer gets 25% of the claimers pool as a separate share, and the named claimers divide the other 75%. On a Default-mode coin with no compounding, Bags’ docs give the example plainly: a 2% fee sends 1% to the platform, 0.25% to the deployer and 0.75% to the named recipients.</p><p>The docs also say there is no public field to choose a different deployer or change its rate. The wallet that pays is the deployer. For a sniper this is useful, because that wallet is the launcher, and it is right there in the record.</p><h3>The partner</h3><p>Apps that launch coins through Bags can take a partner share. It comes out of the platform pool, not out of the claimers’ or the deployer’s part.</p><h3>Collected, claimed and waiting</h3><p>Fees do not reach claimers on every trade. They build up in the pool, and from time to time they are collected into a fee-share authority account tied to the coin. From there each claimer can claim their share whenever they like. The authority account keeps three running totals: everything collected into it, everything the claimers have taken out, and everything the partner has taken out. Whatever is collected and not yet claimed sits in that account’s wrapped SOL balance, in plain view.</p><h3>Two details most people miss</h3><ul><li><strong>The earner list can change.</strong> Bags’ API has an admin endpoint that lets a coin’s fee-share admin change the claimers and their shares. The name on a coin today is not always the name it launched with. When you read a record, read it as it is now, and look at its history if something seems off.</li><li><strong>Some claims are automatic.</strong> With BagsPay, a launcher can set the BagsPay wallet as the only claimer and give a payout handle, such as an X account for X Money, a Cash App or a Venmo account. Every few minutes the BagsPay wallet claims the fees once they reach 0.25 SOL, converts them to USDC, and lists them as a cash order for that handle. So a coin with a steady stream of claims may simply be on BagsPay. That tells you the fees are flowing out. It does not tell you the named person is involved.</li></ul><h3>How often does nobody claim? What the whole program looked like</h3><p>Before building a method on this record, it is worth knowing what it usually says. On November 6, 2026 our team downloaded every fee-share config and every fee-share authority account owned by the Bags Fee Share V2 program and decoded them with the program’s published interface. The full write-up of that count is on our site, so here are only the numbers this guide relies on.</p><ul><li><strong>104,342</strong> fee-share accounts in all, 102,984 of them paired with SOL.</li><li><strong>80,172</strong> had collected any fees at all. <strong>13,285</strong> had collected 1 SOL or more, and 2,705 had collected 10 SOL or more.</li><li>The median SOL-paired coin had collected about <strong>0.12 SOL</strong> in fees over its whole life. Most Bags coins see very little trading.</li><li>Of the 13,285 coins with 1 SOL or more, <strong>5,466, or 41%,</strong> had never seen a claim by any of their named claimers.</li><li>About <strong>31,780 SOL</strong> sat unclaimed across those 13,285 coins, and 27,282 SOL of it was on coins where nobody named had ever claimed.</li></ul><figure><img alt="Chart of how often nobody named on a Bags coin ever claimed its fees, from 13,285 SOL-paired coins that collected 1 SOL or more, read on November 6, 2026. By fees collected: 1 to 10 SOL 43% never claimed, 10 to 100 SOL 34%, 100 SOL or more 22%, all coins 41%. By number of claimers: one claimer 44.5%, two claimers 21%, three or more 12%. About 31,780 SOL sat unclaimed across the 13,285 coins." src="https://cdn-images-1.medium.com/max/1024/1*bp6Fu2R52PSPhZetvl5A7g.png" /></figure><p>Two patterns in that count matter for a watchlist.</p><p><strong>Small coins go unclaimed most.</strong> 43% of coins that collected 1 to 10 SOL had never been claimed. For coins that collected 10 to 100 SOL it was 34%, and for coins that collected 100 SOL or more, 22%. A person who knows they are named on a coin earning a few hundred SOL has every reason to sign in. A coin that earned 3 SOL on its first day and went quiet may never be noticed by the person it names.</p><p><strong>More names, more claims.</strong> 87% of the coins named a single claimer with 100% of the claimers’ share. Those single-claimer coins were never claimed 44.5% of the time. Coins with two claimers, 21%. Coins with three or more, 12%. More names on a coin usually means more people who know about it, and only one of them needs to claim.</p><p>The biggest coins looked different. All five coins with the most fees collected, from about 2,932 SOL to 6,297 SOL, had been claimed, most of them close to in full. The coins that went the distance on Bags nearly always had someone named on them who was paying attention.</p><p>None of that predicts the price of the next coin. What it does is tell you where the information is. Four coins in ten that earned real fees had nobody named on them who ever showed up. A sniper buying the next coin from the same launcher is buying the same pattern.</p><h3>The watchlist method in seven steps</h3><p>A sniper bot fires in the first second of a launch, far too fast for you to read anything. So the reading happens before the launch. You build a short list of launcher wallets whose past coins tell a good story, and you let the bot act on that list. Here is the method we use, step by step.</p><figure><img alt="Flow diagram of seven steps to build a Bags sniper watchlist: 1 start from coins that kept trading, 10 SOL or more in fees; 2 read who is named; 3 check the claims, named earner, partner only, BagsPay or nobody; 4 find the launcher wallet that paid for the config; 5 score it out of five points; 6 note the fee mode; 7 turn it into snipers, one Creator watch per launcher plus a small strict Fresh launch sniper. Repeat weekly." src="https://cdn-images-1.medium.com/max/1024/1*c9rkx1SmE7IVeJgEZByOGg.png" /></figure><h3>Step 1: Start from coins that kept trading, not from new launches</h3><p>New launches are the noisiest place to look. Most of them will earn a fraction of a SOL in fees and then go quiet. Start instead from Bags coins that are at least a few days old and are still trading. Coins that have collected 10 SOL or more in fees are a good floor. In our count only 2,705 coins out of more than 100,000 had reached that level, so this one filter removes almost all of the noise.</p><p>You can find them on any Solana token screener filtered to Bags, or by sorting Bags’ own trending lists. What you want from this step is a list of 20 to 50 coins, each of which proved it could hold attention for more than an afternoon.</p><h3>Step 2: Read who is named on each coin</h3><p>For each coin on your list, read its fee-share config. The Bags app shows each coin’s fee earners and the social accounts behind them, and Bags’ API returns the same roster through its token launch creators endpoint. Write down three things:</p><ul><li>How many claimers are named.</li><li>Whether the main claimer is the person or project the coin is named after.</li><li>Whether the main claimer is the BagsPay wallet. If it is, note the payout handle.</li></ul><p>A coin named after a well-known developer, with that developer’s handle as the only claimer and a launcher wallet that has nothing to do with them, is a tribute coin. That is not wrong, and plenty of tribute coins trade well for a while. But it means the name is the cheapest part of the coin, because anyone can type a handle in seconds.</p><h3>Step 3: Check whether anyone named has ever claimed</h3><p>This is the part that costs something, and so it is the part that tells you something. A claim is a public act. When the named earner signs in and takes the money for the first time, the coin gains a real person connected to it on the record.</p><p>There are two ways to check:</p><ul><li><strong>Bags’ API.</strong> The claim events endpoint lists claims for a coin, page by page or by time range.</li><li><strong>Solscan, no API needed.</strong> Find the coin’s fee-share authority, a public account owned by the Bags Fee Share V2 program, and open its wrapped SOL token account. A balance that only grows, with no claim transactions, means nobody named has claimed. A balance that keeps dropping back near zero means someone is collecting.</li></ul><p>Sort what you find into four groups:</p><ul><li><strong>Named earner claims.</strong> The best sign. Someone on the label is collecting.</li><li><strong>Only the partner claims.</strong> An app that launched the coin is making money from it. The person on the label is not.</li><li><strong>BagsPay claims.</strong> Fees are flowing out automatically to a cash handle. Neutral. It shows the plumbing works, not that the person cares.</li><li><strong>Nobody has ever claimed.</strong> The label has nobody behind it, at least so far.</li></ul><h3>Step 4: Find the launcher</h3><p>The launcher is the wallet that paid for the fee-share config. On current SOL configs it is recorded as the deployer, and it is usually also the wallet that sent the coin’s first transaction. Open that wallet on Solscan and look at what else it has done.</p><p>What you are looking for is a pattern across coins, not one result:</p><ul><li>How many Bags coins has this wallet launched, and how often?</li><li>How many of them kept trading for more than a day?</li><li>On how many did a named earner ever claim?</li><li>Did the wallet buy a big share of its own coins at launch, and did it sell into the first buyers?</li><li>Does it name the same few people again and again?</li></ul><p>In our November 6 count, 146 wallets were named as claimers on 10 or more coins each, together appearing on 4,810 coins. The three busiest were named on 877, 732 and 554 coins and nearly always claimed: across those 2,163 coins only 38 had never seen a claim. A wallet like that is almost certainly a service, such as an app that launches coins for its users. At the other end, one wallet named on 103 coins had never been claimed on 51 of them. Repeat names leave long records, and long records are what a watchlist is made of.</p><h3>Step 5: Score each launcher</h3><p>Keep the scoring simple. You are not trying to predict a price. You are trying to rank a few dozen wallets so you can follow the best five or ten. We use five questions, one point each:</p><ul><li>At least three of its past coins collected 10 SOL or more in fees.</li><li>A named earner claimed on most of those coins.</li><li>Its coins name the person or project behind them, or the launcher itself, not borrowed names.</li><li>It does not usually buy more than about 10% of its own supply at launch.</li><li>It has not sold its launch buy in the first minute on its recent coins.</li></ul><p>A wallet with four or five points goes on the watchlist. Two or three, keep an eye on it. Zero or one, skip it. A launcher whose past coins mostly went quiet will probably launch another that goes quiet. A launcher whose coins kept trading for weeks, with earners who kept collecting, is the one worth a sniper.</p><h3>Step 6: Note the fee mode the launcher uses</h3><p>Launchers tend to stick with one fee mode. Write it down next to each wallet, because it changes every take-profit and stop-loss you will set. The fee math gets its own section below.</p><h3>Step 7: Turn the list into snipers</h3><p>Now the bot can take over. One Creator watch sniper per launcher wallet on your list, each with Bags ticked as its venue. If a wallet on your list is a trader rather than a launcher, a Copy trade sniper is the better fit. And if you still want some exposure to every new launch, run one small Fresh launch sniper with strict filters on the side, sized so that a run of quiet coins does not matter.</p><p>Then repeat steps 1 to 5 once a week. Records keep building, earners start or stop claiming, launchers change habits. A watchlist is a living thing.</p><h3>One coin, read from start to finish</h3><p>Here is what the method looks like on a real coin. The coin is called NOCODE. Its first transaction was on January 18, 2026. Everything below can be checked on Solscan from the links in the sources.</p><figure><img alt="Fee record of one Bags coin, NOCODE, with one claimer holding 100% of the claimers share. About 1,115 SOL in fees collected by November 6, 2026, about 266 SOL claimed by the partner, 0 SOL claimed by the named claimer, and 430.36 SOL still waiting unclaimed on November 16, 2026. Timeline: first transaction January 18, claimer last active February 20, a 0.025 SOL collection on November 14." src="https://cdn-images-1.medium.com/max/1024/1*gFk1LQWdsDYfBZ9yaxAPOQ.png" /></figure><ul><li><strong>Step 1, did it keep trading?</strong> Yes. By November 6 its fee-share authority had collected about 1,115 SOL in fees, which put it among the biggest coins in the whole program.</li><li><strong>Step 2, who is named?</strong> One claimer, with 100% of the claimers’ share.</li><li><strong>Step 3, has anyone named claimed?</strong> No. The claimer’s running total was zero. The partner, on the other hand, had claimed about 266 SOL. The claimer’s wallet held no SOL, and its last transaction was on February 20, 2026.</li><li><strong>What is waiting?</strong> On November 6 the authority’s wrapped SOL account held about 430 SOL. When we checked it again on November 16, 2026, it held 430.36 SOL, still unclaimed. The account was still active: on November 14 a routine collection moved another 0.025 SOL of pool fees into it.</li></ul><p>So the coin earned hundreds of SOL for the person it was set up to pay, an app collected its partner share the whole time, and the named earner has not touched any of it in nine months. We do not know who the claimer is. An untouched embedded wallet fits a social account whose owner has never signed in, but the chain alone cannot prove that.</p><p>What would this coin have told a sniper? Not that it would fail. It clearly did not, at least for a long while. It would have told them that the coin’s story did not include the person on its label. If its launcher had launched another coin the next week, that new coin would have been worth judging on the launcher’s record, not on the name it carried. That is the whole idea of the watchlist.</p><p>NOCODE is not unusual in shape, only in size. Among the ten never-claimed coins with the most fees waiting on November 6, nine named a single claimer.</p><h3>What does it cost to trade a Bags coin? The fee math</h3><p>The same fee that pays the earners is paid by you, on every buy and every sell. It is easy to forget inside a sniper, because it is not a token tax and most screens do not show it as one. Bags’ docs list seven SOL fee modes:</p><ul><li><strong>Default.</strong> 2% on the curve, 2% after graduation, with a quarter of the post-graduation fee added back to the pool.</li><li><strong>Low Float.</strong> 2% on the curve. After graduation the fee starts at 2% and falls as the market cap grows, down to a 0.5% floor at roughly 25 times the market cap at graduation. 96% of supply is locked. Graduates at 80 SOL.</li><li><strong>Drop Mode.</strong> Same fees as Default, with 85% of supply locked. Graduates at about 100 SOL.</li><li><strong>Price Mode.</strong> Same fees as Default, with a total supply of 1,000 tokens.</li><li><strong>About 0% Fee Mode.</strong> 1% on the curve, 0.25% after graduation.</li><li><strong>1% Fee Mode.</strong> 0.25% on the curve, 1% after graduation, with half of the post-graduation fee added back to the pool.</li><li><strong>Paper Hand Tax Mode.</strong> 10% on the curve and 10% after graduation.</li></ul><p>A round trip pays the fee twice. The simple way to see what that means is to ask how far the price has to rise before you are back to even. If the buy fee is b and the sell fee is s, you need the price to rise by 1 divided by ((1 minus b) times (1 minus s)), minus 1. Here is what that gives for a buy and a sell both on the curve, before network fees, priority fees, tips and slippage:</p><ul><li><strong>Paper Hand Tax Mode:</strong> the price has to rise about <strong>23.5%</strong> just to break even.</li><li><strong>Default, Low Float, Drop and Price:</strong> about <strong>4.1%</strong>.</li><li><strong>About 0% Fee Mode:</strong> about <strong>2.0%</strong>.</li><li><strong>1% Fee Mode:</strong> about <strong>0.5%</strong>.</li></ul><figure><img alt="Chart of what a buy and a sell on the Bags bonding curve cost by fee mode, before network fee, tip and slippage. Price rise needed to break even: Paper Hand Tax 10% each way 23.5%; Default, Low Float, Drop and Price 2% each way 4.1%; About 0% Fee Mode 1% each way 2.0%; 1% Fee Mode 0.25% each way 0.5%. Kept from a 40% move: 13.4%, 34.5%, 37.2% and 39.3%. Worked example, not trade results." src="https://cdn-images-1.medium.com/max/1024/1*ACOXLIeJaUXjluQham15UA.png" /></figure><p>Turn it around and look at a good trade. Say the price rises 40% between your buy and your sell, both on the curve. In 1% Fee Mode you keep about 39.3% of that move. In the About 0% Fee Mode, about 37.2%. In Default, about 34.5%. In Paper Hand Tax Mode, about 13.4%. Same coin chart, same timing, and the mode alone takes away most of the gain on the expensive one. These are worked examples of the math, not results from real trades.</p><p>What this means for settings:</p><ul><li><strong>On Default-mode coins,</strong> put the first take-profit at +30% or more. A +10% take-profit is closer to +6% in your wallet.</li><li><strong>On Paper Hand Tax coins,</strong> quick flips almost never pay. Either treat them as long holds with wide targets, or block the launchers who use that mode.</li><li><strong>On the cheap modes,</strong> tighter targets are fine on the curve, but check what happens after graduation. The 1% Fee Mode is cheap on the curve and four times as expensive after it.</li><li><strong>Count the fee in your stop-loss too.</strong> A 25% stop-loss on a Default-mode coin loses closer to 28% after both fees.</li></ul><p>There is one small point in Bags’ favour for migration snipers. In Default mode a quarter of the post-graduation fee is added back into the pool’s liquidity, so a pool that keeps trading gets deeper on its own. That makes it a little easier to get a larger position out later.</p><p>If you have read <a href="https://medium.com/@dexsnipecom/orca-sniper-bot-three-buyers-one-whirlpool-and-the-fee-that-decided-who-made-money-dae478d23913">our Orca guide</a>, the lesson is the same one in a different shape. There the fee moved with the price. On Bags it is fixed at launch, which is easier, because you can read it before you buy.</p><h3>Bags sniper bot settings that fit the watchlist</h3><p>Two snipers for two jobs. These are starting points, not promises. Adjust them to your own budget and to what your watchlist tells you.</p><h3>Creator watch sniper, one per launcher on your list</h3><ul><li><strong>Trigger:</strong> Creator watch, the launcher’s wallet, venue Bags.</li><li><strong>Buy size:</strong> 0.2 to 0.5 SOL. These are your best-researched entries, so they get the larger size.</li><li><strong>Slippage cap:</strong> 20%.</li><li><strong>Priority fee:</strong> 0.001 SOL. <strong>Jito tip:</strong> 0.002 SOL.</li><li><strong>Max dev holding:</strong> 10%.</li><li><strong>Skip bundled launches:</strong> on.</li><li><strong>Max top 10 holders:</strong> 35%.</li><li><strong>Take-profit:</strong> sell 50% at +50%, the rest at +150%.</li><li><strong>Stop-loss:</strong> 30%. <strong>Timer:</strong> 60 minutes. <strong>Dev-sell exit:</strong> on.</li></ul><h3>Fresh launch sniper, small and strict, on the side</h3><ul><li><strong>Trigger:</strong> Fresh launch, venue Bags.</li><li><strong>Buy size:</strong> 0.05 to 0.1 SOL.</li><li><strong>Slippage cap:</strong> 15%.</li><li><strong>Priority fee:</strong> 0.0005 SOL. <strong>Jito tip:</strong> 0.001 SOL.</li><li><strong>Max dev holding:</strong> 5%.</li><li><strong>Skip bundled launches:</strong> on. <strong>Require socials:</strong> on.</li><li><strong>Max top 10 holders:</strong> 25%.</li><li><strong>Blocklist words:</strong> the names of people or projects your watchlist shows are often launched in tribute without their involvement.</li><li><strong>Take-profit:</strong> sell 50% at +30%, the rest at +80%.</li><li><strong>Stop-loss:</strong> 25%. <strong>Timer:</strong> 15 minutes. <strong>Dev-sell exit:</strong> on.</li></ul><h3>Why the timers are short</h3><p>The median Bags coin collects about 0.12 SOL in fees over its whole life. At 2%, that is only about 6 SOL of trading in its whole life. Most coins go quiet quickly, and a position in a quiet coin is SOL doing nothing. A timer frees it. Your Creator watch snipers get a longer timer because those launchers have earned a little more patience.</p><h3>Why the socials check is only a floor</h3><p>Tribute coins often borrow the named person’s links. A coin can pass a socials check and still have nobody behind it. Keep the check on to drop the laziest launches, but do not read a link as proof of anything. For a deeper look at which filters actually separate launches, <a href="https://medium.com/@dexsnipecom/pump-fun-sniper-bot-filters-we-checked-807-new-coins-to-see-which-rules-work-5c7a932f9e9e">our filter test on 807 new coins</a> goes through them one by one. It was run on Pump.fun, but most of its lessons carry over.</p><h3>Migration is a separate trade</h3><p>After a coin graduates to its DAMM v2 pool, the fee may change with the mode and the liquidity is very different. If you want to trade graduations, run a Migration sniper with its own settings instead of stretching a launch sniper to cover both.</p><h3>Six mistakes we see with Bags snipers</h3><ul><li><strong>Treating the name as the team.</strong> The name on a Bags coin is the one thing that costs nothing to add. Read the claim record.</li><li><strong>Reading one coin instead of a launcher.</strong> One coin is a story. Ten coins from the same wallet are a record.</li><li><strong>Reading BagsPay claims as involvement.</strong> Automatic claims show the fees are flowing. They do not show the person cares.</li><li><strong>Forgetting the fee mode.</strong> A +15% take-profit on a Paper Hand Tax coin is a loss.</li><li><strong>Never refreshing the list.</strong> Earners can change, admins can update configs, and launchers change habits. Re-score weekly.</li><li><strong>Sizing the Fresh launch sniper like the watchlist snipers.</strong> The side sniper should be small enough that ten quiet coins in a row do not hurt.</li></ul><h3>What this record cannot tell you</h3><ul><li><strong>It is not a price signal.</strong> We did not test whether unclaimed coins trade worse. The record tells you who is behind a coin, not where it goes.</li><li><strong>Unclaimed is not dishonest.</strong> A person may not need the money yet, or may claim through a payout service. Read it as evidence about involvement, not about honesty.</li><li><strong>Snapshots age.</strong> A coin unclaimed today may be claimed tomorrow. Our program-wide count is from November 6, 2026.</li><li><strong>Collected is not earned.</strong> Fees still sitting in a pool are not in the authority account yet, so recent coins look smaller than they will.</li><li><strong>Older programs are not included.</strong> Coins set up under the legacy Fee Share V1 program are not in our count.</li><li><strong>The launcher’s own cut runs separately.</strong> On newer configs the deployer collects through its own flow, so a launcher may be collecting on a coin where the named earners never did. Our never-claimed count is about the named earners only.</li></ul><h3>How to set up a Bags sniper bot in DexSnipe</h3><p>DexSnipe is our own Solana sniper bot, so this is how we would run the watchlist in it. Any bot that has a Creator watch trigger, holder checks and server-side exits can do the same job.</p><ul><li><strong>1. Open the panel and click New sniper.</strong> Choose Creator watch, paste a launcher wallet from your list, and tick Bags. Add one sniper for each launcher you follow.</li><li><strong>2. Set the buy.</strong> Buy size, a slippage cap of about 20%, and a modest priority fee and Jito tip.</li><li><strong>3. Set the filters.</strong> Max dev holding 10%, skip bundled launches on, max top 10 holders 35%.</li><li><strong>4. Build exits wide enough for the fee.</strong> Up to four take-profit steps, a stop-loss, a trailing stop if you want one, a timer and the dev-sell exit.</li><li><strong>5. Add the side sniper.</strong> One Fresh launch sniper on Bags with the strict settings above.</li><li><strong>6. Connect your wallet and save.</strong> Snipers are stored on your wallet address. Your main wallet is only used to sign in.</li><li><strong>7. Pick a plan, add a sniper wallet and arm it.</strong> Plans are 1 SOL for 7 days, 3 SOL for 30 days or 5 SOL for 90 days, with no fee per trade. Add the key of a separate funded wallet for the snipers to trade from, arm them, and close the tab. Exits keep running on our side.</li></ul><p>A plan covers all 12 venues DexSnipe watches and all six triggers, up to 12 snipers and 5 sniper wallets. On a Bags trade you pay the network fee, your priority fee and tip, and the coin’s own Bags trading fee. DexSnipe does not add a cut on top.</p><h3>Frequently asked questions</h3><h3>What is the best Bags sniper bot?</h3><p>The best one is the one that lets you act on what you know about a launcher. Look for support for Bags launches and migrations on Meteora’s bonding curve and DAMM v2 pools, a Creator watch or wallet-follow trigger, holder checks before the buy, exits that run on a server when you are offline, and a flat price. Bags already charges a fee on every trade, so a bot that adds its own cut on top makes every target harder to reach.</p><h3>Who earns the trading fees on a Bags coin?</h3><p>The claimers named in the coin’s fee-share config, up to 100 of them, plus Bags’ platform pool, an optional partner and, on newer SOL configs, the wallet that paid for the launch, which takes 25% of the claimers’ side as a separate deployer share.</p><h3>Can someone launch a Bags coin in my name without asking me?</h3><p>Yes. A launcher can name a social handle as the fee earner, and Bags resolves it to an embedded wallet. The named person does not need to have a wallet or agree. They can sign in later and claim what has built up.</p><h3>How do I check if a Bags coin’s fees have been claimed?</h3><p>Find the coin’s fee-share authority, an account owned by the Bags Fee Share V2 program, on Solscan and open its wrapped SOL token account. A balance that keeps growing with no claim transactions means nobody named has claimed. Bags’ API also has a claim events endpoint for each coin.</p><h3>Does a claim mean the coin is safe?</h3><p>No. It means someone named on the coin is collecting. That is a sign of involvement, not a promise about price, and BagsPay claims are automatic, so they do not even show that.</p><h3>What fee does a Bags coin charge?</h3><p>It depends on the mode picked at launch. The Default mode charges 2% on every trade. Other modes range from 0.25% to 10%, and some change after graduation. Check the mode before you set a take-profit.</p><h3>How much SOL does a Bags coin need to graduate?</h3><p>Most fee modes graduate at 85 SOL raised on the curve. Low Float graduates at 80 SOL, and Drop Mode at about 100 SOL. After that the coin trades in a Meteora DAMM v2 pool.</p><h3>Should I use Fresh launch or Creator watch on Bags?</h3><p>Mostly Creator watch. On Bags, who launched a coin and whether the people on it ever show up carries more information than on most launchpads. Fresh launch ignores that. Keep a small Fresh launch sniper if you want some exposure to everything, but put the larger size behind launchers you have checked.</p><h3>Does DexSnipe take a fee on Bags trades?</h3><p>No. DexSnipe is a flat plan: 1 SOL for 7 days, 3 SOL for 30 days or 5 SOL for 90 days. You pay the network fee, your own priority fee and tip, and the coin’s Bags trading fee.</p><h3>How we collected the data</h3><p>The program-wide numbers come from our read of the Bags Fee Share V2 program on November 6, 2026. We listed every account owned by the program with the discriminator of a fee-share config or a fee-share authority, decoded them with the layouts in the program’s published interface, and matched each authority to its config. We kept configs whose quote asset is wrapped SOL, and for every coin whose authority had collected 1 SOL or more we read the claimer list, the shares and the balance of the authority’s wrapped SOL token account. “Never claimed” means the authority’s total claimed by users was zero.</p><p>For the example coin, we read the authority’s balance and transactions and the claimer’s balance and history directly from a public Solana RPC node on November 6 and again on November 16, 2026. The November 14 transaction was decoded from its logs and token balances. Fee modes, graduation amounts, the deployer share, the admin update endpoint and BagsPay come from Bags’ documentation, read on November 16, 2026; they can change, so check a coin’s own mode before trading it. The break-even and kept-gain figures are worked examples of the fee formula, not trade results. DexSnipe triggers, filters and exits are taken from the product itself.</p><h3>Sources and further reading</h3><ul><li>Bags docs, Program IDs (Fee Share V2, Meteora DBC and DAMM v2, and where the interfaces are published): <a href="https://docs.bags.fm/principles/program-ids">docs.bags.fm</a></li><li>Bags docs, Customize Token Fees (the seven SOL fee modes, the fee split and graduation amounts): <a href="https://docs.bags.fm/how-to-guides/customize-token-fees">docs.bags.fm</a></li><li>Bags docs, Deployer Fees (the 25% deployer share of the claimers pool): <a href="https://docs.bags.fm/how-to-guides/deployer-fees">docs.bags.fm</a></li><li>Bags docs, Create Fee Share Config (up to 100 claimers, basis points summing to 10,000): <a href="https://docs.bags.fm/api-reference/create-fee-share-configuration">docs.bags.fm</a></li><li>Bags docs, Create Fee Share Admin Update Config (changing claimers after launch): <a href="https://docs.bags.fm/api-reference/create-fee-share-admin-update-config">docs.bags.fm</a></li><li>Bags docs, Pay Fees to X Money (BagsPay automatic claims and cash payouts): <a href="https://docs.bags.fm/how-to-guides/pay-fees-to-x-money">docs.bags.fm</a></li><li>Bags docs, Get Token Claim Events: <a href="https://docs.bags.fm/how-to-guides/get-token-claim-events">docs.bags.fm</a></li><li>Bags SDK interfaces on GitHub (the Fee Share V2 layout we decoded): <a href="https://github.com/bagsfm/bags-sdk/tree/main/src/idl">github.com/bagsfm</a></li><li>Bags Fee Share V2 program on Solscan: <a href="https://solscan.io/account/FEE2tBhCKAt7shrod19QttSVREUYPiyMzoku1mL1gqVK">solscan.io</a></li><li>The example coin, NOCODE, on Solscan: <a href="https://solscan.io/token/79VXvEt7xSPhvQggeJ75UVqD41yXJwBS44zrstreBAGS">solscan.io</a></li><li>Its fee-share authority on Solscan (the 430.36 SOL waiting): <a href="https://solscan.io/account/5q2GSwqTD5CisdHbbzFtsBBS2KPt6f68em5B1c6HxqgX">solscan.io</a></li></ul><p>Docs were checked in November 2026. Nothing here is financial advice. Meme coins are very risky, and most new tokens lose most of their value. Knowing who is behind a coin improves your choices; it does not make new coins safe. Only trade with money you can afford to lose.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=421549cc7cbe" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Orca Sniper Bot: Three Buyers, One Whirlpool, and the Fee That Decided Who Made Money]]></title>
            <link>https://medium.com/@dexsnipecom/orca-sniper-bot-three-buyers-one-whirlpool-and-the-fee-that-decided-who-made-money-dae478d23913?source=rss-2654b58e3d1e------2</link>
            <guid isPermaLink="false">https://medium.com/p/dae478d23913</guid>
            <category><![CDATA[memecoins]]></category>
            <category><![CDATA[trading-bot]]></category>
            <category><![CDATA[solana-blockchain]]></category>
            <category><![CDATA[cryptocurrency]]></category>
            <category><![CDATA[crypto-trading]]></category>
            <dc:creator><![CDATA[DexSnipe]]></dc:creator>
            <pubDate>Sun, 15 Nov 2026 05:50:54 GMT</pubDate>
            <atom:updated>2026-11-15T05:50:54.154Z</atom:updated>
            <content:encoded><![CDATA[<h3>One new Orca pool, three buyers a few seconds apart, and a trading fee that went from 1% to 10% in under two minutes. What each one paid, what it did to their profit, and how to set up an Orca sniper bot so you end up in the right seat.</h3><p>On October 29, 2026 a coin called FOMOEYES got a new pool on Orca. The pool had a 1% fee on the label. The first person to buy paid exactly that. A buyer who came in 34 seconds later paid 9.19%. A group of buyers about a minute and a half in paid 10%, every one of them, and so did the people who sold to them. Nobody changed a setting in between. The fee moved because the price moved. That is how Orca’s adaptive fee pools work, and if you run an Orca sniper bot it is the most important thing to understand about the venue. This guide follows that one pool from three seats: the first buyer, the chaser and the wave. Then it shows what the same pattern looked like across 76 new pools, how to read an Orca pool before your bot fires, and the settings that put you in the first seat instead of the third.</p><figure><a href="https://www.dexsnipe.com/"><img alt="DexSnipe banner: Snipe Orca on your own rules. New pool, creator watch and copy trade snipers with liquidity and holder checks, exits run on the server. The button opens https://www.dexsnipe.com/ and an illustration of the panel shows Orca snipers with min liquidity 10 SOL, dev cap 15% and top 10 holders max 30%." src="https://cdn-images-1.medium.com/max/1024/1*EcmFLfOiwHBIh0QhVQ51FQ.png" /></a><figcaption>DexSnipe is our own Solana sniper bot service.</figcaption></figure><h3>The short version</h3><ul><li><strong>The fee on many new Orca pools is not fixed.</strong> Orca’s adaptive fee pools charge a base fee when trading is calm and add more when the price moves fast. The program caps the total at 10%.</li><li><strong>The first buyer usually pays the least.</strong> A calm pool has no volatility yet, so the first trade pays the base fee or close to it. In the FOMOEYES pool the first buy paid 1.00%.</li><li><strong>The crowd behind them pays the most.</strong> Across 76 new adaptive pools, buys in the first ten seconds paid a median fee of 2.0%. Buys between minute one and minute five paid a median of 5.05%, and more than half of them paid 5% or more.</li><li><strong>The fee hits your exit too.</strong> Selling into a fast move is the most expensive time to sell. A take-profit that fires in the middle of a buying wave can lose close to a tenth of the sale.</li><li><strong>So a good Orca sniper does two things.</strong> It lands early or skips the pool, and it sets targets that still make sense after paying up to 10% on the way in and 10% on the way out.</li></ul><h3>First, what Orca is and what an Orca sniper bot buys</h3><p>Orca is not a launchpad. It is a decentralized exchange on Solana, and its pools are called Whirlpools. A Whirlpool is a concentrated liquidity pool: the people who add liquidity can choose a price range, and their liquidity only works while the price is inside that range. Anyone can create a Whirlpool for any pair of tokens.</p><p>That last part is why snipers care. Coins from every launchpad end up with Orca pools. Pump.fun coins, LetsBonk coins, Jupiter Studio coins and Bags coins all get them, created by their teams, by market makers, by launch services or by anyone else who wants a pool. Orca’s own launchpad, Wavebreak, graduates coins straight into Whirlpools. So an Orca sniper bot is less about one launchpad’s rules and more about one exchange’s rules.</p><p>A sniper on Orca does a simple job. It watches the Whirlpools program for new pools. When a pool gets real liquidity, it checks the coin, buys, and then sells on rules you set. A second style, creator watch, narrows that down to pools created by wallets you choose. A third, copy trading, follows wallets you choose and buys what they buy.</p><p>For years the fee on a Whirlpool was simple. Every pool had a fee tier, like 0.3% or 1%, and that was the fee on every swap. Orca’s docs list tiers from 0.01% up to 2%, and the chain shows a few hundred pools on a 6% tier as well. In 2025 Orca added adaptive fee pools, and the tier on those pools stopped being the fee. It became the floor.</p><h3>The pool: FOMOEYES, October 29, 2026</h3><p>The FOMOEYES pool is a good example because it is so ordinary. It is an adaptive splash pool with a 1% base fee and the most common adaptive settings on Orca, the same settings used by 14,892 other pools.</p><p>A splash pool is Orca’s pool type for token launches. It is full range, with a very wide tick spacing (32896) and a 1% fee tier, so the liquidity covers every price and the pool behaves like a classic constant product pool. Orca’s developer docs say splash pools are best for token launches and for liquidity providers who do not want to manage a range. Out of 167,036 Whirlpools on chain when we read them on November 6, 2026, 15,639 were splash pools, and 8,367 of those used adaptive fees.</p><p>Here are the first two and a half minutes of trading in the FOMOEYES pool, with the fee each trade actually paid. The fee comes from the trade event that Orca’s program writes for every swap, which records the amount in and the fee taken.</p><ul><li><strong>0 seconds:</strong> buy, fee 1.00%</li><li><strong>23 seconds:</strong> buy, fee 3.63%</li><li><strong>34 seconds:</strong> buy, fee 9.19%</li><li><strong>46 seconds:</strong> sell, fee 4.25%</li><li><strong>59 seconds:</strong> buy, fee 1.25%</li><li><strong>65 seconds:</strong> buy, fee 6.28%</li><li><strong>78 to 92 seconds:</strong> five buys, fee 10.00% each</li><li><strong>100 seconds:</strong> sell, fee 8.82%</li><li><strong>100 to 126 seconds:</strong> eight buys and one sell, fee 10.00% each</li><li><strong>131 seconds:</strong> two buys, fee 4.46%</li></ul><figure><img alt="Chart of the fee paid on every trade in the first 131 seconds of the FOMOEYES Orca adaptive splash pool, base fee 1%. The first buy at 0 s paid 1.00%, a buy at 23 s paid 3.63%, the chaser at 34 s paid 9.19%, a sell at 46 s paid 4.25%, the fee fell back to 1.25% at 59 s, then 13 buys between 78 and 126 s paid the 10% cap, with sells at 8.82% and 10%. Two buys at 131 s paid 4.46%." src="https://cdn-images-1.medium.com/max/1024/1*COVm9Yw_13BOuYfirewyAA.png" /></figure><p>Read that list twice and the whole venue makes sense. The fee started at the base. It jumped as soon as buys started pushing the price. It dropped back during a short pause around the one minute mark. Then a wave of buys pushed it to the 10% cap, and it stayed there for most of a minute. Now let us sit in three different seats.</p><h3>Seat one: the first buyer</h3><p>The first buy in the pool paid 1.00%. That is the base fee, and it is the lowest fee anyone paid in the whole opening.</p><p>This is the part of Orca that surprises people who have sniped other venues. On many launch venues, being first is the most expensive thing you can do. On Meteora, many launches open with an anti-sniper fee of 50% or more that falls over the first minute. On Jupiter Studio, launches with anti-sniping on charge 99% at first. On Believe, a buy in the first seconds paid close to 50%. Those fees are written into the launch config and depend only on time. They tax speed on purpose.</p><p>Orca’s adaptive fee works the other way. It depends on how much the price has moved recently, and a pool that has not traded yet has not moved. So the first buyer walks into a calm pool and pays the base fee. In our wider sample, the 67 buys that landed in the very first second of trading paid a median fee of 1.5%.</p><p>That gives the first seat two real advantages. The fee on the way in is low. And the first buyer is holding tokens when everyone else starts pushing the price up, which is the whole point of sniping.</p><p>The first seat still has one problem, and it comes later. When the first buyer wants to sell into the rush, they are selling into the most expensive moment in the pool. We will come back to that.</p><h3>Seat two: the chaser</h3><p>The third buy came 34 seconds after the first, and it paid 9.19%. Same pool, same settings, same coin. More than nine times the first buyer’s fee.</p><p>What happened in between is simple. Two buys pushed the price up across several groups of ticks, and every time the price crosses tick groups, the pool’s stored volatility grows. The adaptive part of the fee grows with the square of that number, so it climbs fast once it gets going. By the time the chaser arrived, the pool was charging close to the cap.</p><p>The chaser is the bot or person who saw the pool a little late and fired anyway. On Orca this is the worst position to buy from. You pay a high fee on the way in, and you buy after the price has already moved, so you need an even bigger move to make money.</p><p>Notice what happened right after. At 46 seconds someone sold and paid 4.25%. At 59 seconds a buy paid only 1.25%. There was a short pause, and the fee fell back almost to the base. That pause is important. It means a sniper that misses the open is not forced to pay 9% or 10%. If it waits for a quiet moment, it can sometimes buy close to the base fee again. A bot that fires the moment it sees the pool, no matter what the pool is doing, gets whatever fee the last half minute left behind.</p><h3>Seat three: the wave</h3><p>From 78 to 126 seconds, thirteen buys went into the FOMOEYES pool, and every one of them paid the full 10%. That is the cap built into Orca’s program. The fee cannot go higher than that.</p><p>The wave is the group of buyers who arrive once a coin starts to look busy. They see the price going up, the trade count going up and the chart going green, and they pile in together. On a fixed-fee pool that would cost them 1% each. On an adaptive pool, their own buying keeps the volatility high, and the fee sits at the cap until the buying stops.</p><p>The wave pays the most, and it also buys at the highest prices. Out of the three seats, it is the one with the least room for profit.</p><h3>The sellers in the middle</h3><p>Now look at the sells in the list. At 46 seconds a sell paid 4.25%. At 100 seconds, in the middle of the wave, a sell paid 8.82%. Between 100 and 126 seconds, a sell paid 10%.</p><p>Those sellers are probably early buyers taking profit, maybe even the first buyer. They did the right thing by selling into strength. But strength on an adaptive pool means a fast-moving price, and a fast-moving price is exactly when the fee is highest. The fee applies the same way to buys and sells.</p><p>This is the part most Orca guides miss. The first seat gets a cheap entry, but the exit can still cost up to 10% if the take-profit fires into the rush.</p><h3>The math: what each seat keeps from the same move</h3><p>Here is a simple way to see what the fee does to a trade. The numbers below are an illustration. We do not know each FOMOEYES trader’s price or size, so we use the real fees from the list and an example move: the price rises 40% from where you bought, and you sell.</p><p>The formula is short. You pay the buy fee on the way in, so you get fewer tokens. You pay the sell fee on the way out, so you keep less SOL. What you end with is your starting SOL times (1 minus the buy fee) times (1 plus the move) times (1 minus the sell fee).</p><ul><li><strong>First buyer, sells in a calm moment (1% in, 1% out):</strong> 0.99 x 1.40 x 0.99 = 1.372. You keep about +37.2% out of the 40% move.</li><li><strong>First buyer, sells into the rush (1% in, 10% out):</strong> 0.99 x 1.40 x 0.90 = 1.247. About +24.7%. The exit fee took a third of the profit.</li><li><strong>Chaser, sells calm (9.19% in, 1% out):</strong> 0.9081 x 1.40 x 0.99 = 1.259. About +25.9%.</li><li><strong>Chaser, sells into the rush (9.19% in, 10% out):</strong> 0.9081 x 1.40 x 0.90 = 1.144. About +14.4%.</li><li><strong>Wave buyer, sells into the rush (10% in, 10% out):</strong> 0.90 x 1.40 x 0.90 = 1.134. About +13.4%. Nearly two thirds of the move went to fees.</li></ul><p>Now turn it around and ask a different question: how far does the price need to rise just to get your money back?</p><ul><li><strong>1% in and 1% out:</strong> the price needs to rise about 2.0% to break even.</li><li><strong>1% in and 10% out:</strong> about 12.2%.</li><li><strong>9.19% in and 10% out:</strong> about 22.4%.</li><li><strong>10% in and 10% out:</strong> about 23.5%.</li></ul><figure><img alt="Bar chart of the price rise needed to break even and the share of a 40% move kept, using real FOMOEYES fees. First buyer, calm exit, 1% in and 1% out: needs 2.0%, keeps +37.2%. First buyer exiting in the rush, 1% in and 10% out: needs 12.2%, keeps +24.7%. Chaser, 9.19% in and 10% out: needs 22.4%, keeps +14.4%. Wave buyer, 10% in and 10% out: needs 23.5%, keeps +13.4%." src="https://cdn-images-1.medium.com/max/1024/1*SoTvK5FPhzi3HYNGxqO_eA.png" /></figure><p>That is the whole story in four lines. The first buyer who also exits in a calm moment needs a 2% move to break even. The wave buyer who exits into the rush needs more than 23%. Same pool, same coin, same two minutes. The seat decides it.</p><p>It also explains a common complaint from people who start sniping on Orca: “my take-profit hit, but I barely made anything.” A +30% target that fires with 2% in and 10% out leaves about +14.7%. A +30% target with 10% in and 10% out leaves about +5.3%. The target was hit. The fees ate it.</p><h3>Was FOMOEYES unusual? What 76 new pools show</h3><p>One pool is a story, not a rule. To check whether FOMOEYES was normal, we took 76 adaptive pools paired with SOL, created at even intervals through 2026, and read the first 100 transactions of each one. For every swap, we divided the fee by the amount in to get the fee that trade actually paid, then grouped trades by time since the pool’s first trade.</p><p>Here is what buys paid:</p><ul><li><strong>0 to 10 seconds:</strong> 88 buys, median fee 2.00%, 20% of buys paid 5% or more</li><li><strong>10 to 60 seconds:</strong> 59 buys, median fee 3.05%, 37% paid 5% or more</li><li><strong>1 to 5 minutes:</strong> 203 buys, median fee 5.05%, 53% paid 5% or more</li><li><strong>5 to 60 minutes:</strong> 656 buys, median fee 4.47%, 48% paid 5% or more</li><li><strong>After an hour:</strong> 597 buys, median fee 1.85%, 22% paid 5% or more</li></ul><p>And here is what sells paid in the same pools:</p><ul><li><strong>0 to 10 seconds:</strong> 45 sells, median fee 1.45%, 13% paid 5% or more</li><li><strong>10 to 60 seconds:</strong> 46 sells, median fee 2.31%, 28% paid 5% or more</li><li><strong>1 to 5 minutes:</strong> 104 sells, median fee 2.62%, 38% paid 5% or more</li><li><strong>5 to 60 minutes:</strong> 297 sells, median fee 3.60%, 42% paid 5% or more</li><li><strong>After an hour:</strong> 496 sells, median fee 1.63%, 20% paid 5% or more</li></ul><p>The pattern matches FOMOEYES closely. The first seconds were the cheapest time to buy. The fee climbed as trading picked up and peaked between minutes one and five, where the median buy paid about five times the 1% base most of these pools use. Across all trades in that window, 29% paid the full 10%. 34 of the 76 pools charged close to 10% on at least one trade in their opening, and 49 charged 5% or more. After the first hour, things calmed down and the median buy paid 1.85%.</p><p>So FOMOEYES was not an odd case. It was a clear picture of what usually happens.</p><h3>How the adaptive fee actually works</h3><p>You do not need the math to run a sniper, but knowing the moving parts helps you read a pool. This comes from Orca’s docs and from the open-source Whirlpools program on GitHub.</p><p>Every adaptive pool has an extra account called an oracle. A pool without an oracle charges a fixed fee. A pool with one charges a base fee plus an adaptive part. The oracle stores the adaptive settings and a running measure of recent price movement, called the volatility accumulator.</p><ul><li><strong>When a swap moves the price across groups of ticks, the accumulator grows.</strong> Bigger moves make it grow more.</li><li><strong>The adaptive part of the fee follows the square of the accumulator,</strong> scaled by a control setting chosen when the pool was made. That is why the fee climbs slowly at first and then very fast.</li><li><strong>Base fee plus adaptive part is capped at 10%.</strong> This limit is written in the program.</li></ul><p>Three timers decide how fast the fee comes back down. On almost every adaptive pool we read, they were set the same way:</p><ul><li><strong>Filter period, 30 seconds.</strong> Swaps within 30 seconds of the last reference update build on the same volatility. A burst of buys keeps stacking.</li><li><strong>Decay period, 600 seconds.</strong> Between 30 seconds and ten minutes after the last update, the stored volatility is cut, by half on almost every pool we read.</li><li><strong>After ten minutes of quiet,</strong> the stored volatility goes back to zero and the fee is back to base.</li></ul><p>Put that next to the FOMOEYES list and you can see every step. The first buy hit a calm pool. Buys inside the same 30 seconds stacked. A short gap let the fee fall back to 1.25%. Then a long burst of buys kept the accumulator high and the fee stuck at 10% until the burst ended.</p><p>How steep the fee gets depends on the settings. On the most common setup, used by 14,892 pools and almost all with a 1% base fee, the adaptive part can add far more than 9% once the accumulator is full, so the 10% cap is what stops it. Out of 25,674 adaptive pools, 22,468 have settings steep enough to reach the cap. Most of the rest top out at between 0.7% and 7.4% of extra fee.</p><h3>How common are adaptive pools?</h3><p>On November 6, 2026 we downloaded every Whirlpool and every oracle account of the Whirlpools program. Here is what the chain showed:</p><ul><li><strong>Whirlpools:</strong> 167,036</li><li><strong>Adaptive fee pools:</strong> 25,674</li><li><strong>Adaptive pools paired with SOL:</strong> 10,579</li><li><strong>Adaptive pools that can reach the 10% cap:</strong> 22,468</li><li><strong>Adaptive pools for coins whose mint ends in “pump”:</strong> 5,466</li><li><strong>Fixed-fee pools for coins whose mint ends in “pump”:</strong> 3,375</li></ul><p>Adaptive pools are a minority of all Whirlpools, but not among memecoin pools. For coins with the “pump” suffix, which are almost always Pump.fun coins, adaptive pools outnumber fixed-fee ones. If your Orca sniper mostly buys launchpad coins, most of its fills will be in pools whose fee moves.</p><p>New adaptive pools also keep coming at a steady pace. We dated a sample of 500 adaptive pools by their first transaction. The earliest was from May 2025, and from July 2025 on every month contributed between 22 and 40 of the 500. Scaled to all adaptive pools, that is somewhere between about 1,100 and 2,000 new adaptive pools a month. Fixed-fee pools are still created too, and in larger numbers. A sniper on Orca sees both kinds side by side, so it should never assume which one it is looking at.</p><h3>Two more things a new Orca pool can do to your sniper</h3><p>The fee is the big one, but two other details catch snipers out on Orca.</p><p><strong>New pools start empty.</strong> We read the first transactions of 80 adaptive SOL pools created in 2026. Not one received liquidity or a swap in the transaction that created it. Every pool was created in one transaction and funded in a later one. The median first swap came about a minute after creation. A tenth traded within five seconds, and a tenth waited close to an hour. A sniper that fires the moment a pool is created is trying to buy into an empty pool. That is why the minimum liquidity filter matters more on Orca than almost anywhere else.</p><p>The same 80 pools were made by 68 different wallets, and one wallet made nine of them. A wallet that creates pools for coin after coin is often a launch service or a market maker, and its pools tend to follow a pattern. That is useful for creator watch: you can follow a wallet whose pools tend to go somewhere, or block one whose pools never do.</p><p><strong>Some pools tell you exactly when trading starts.</strong> An adaptive pool’s oracle can hold a trade enable time. If it is set, the program refuses every swap until that moment. Orca’s code allows it to be up to 72 hours after the pool is created. 85 adaptive pools had one set. When we checked each one’s first transaction against its start time, trading opened exactly 30 seconds after creation every time. 69 of the 85 are for coins whose mints end in “wave”: graduations from Wavebreak, Orca’s own launchpad, which SolanaFloor described at launch as built around anti-bot protection.</p><p>A scheduled start turns the race into a timing problem. The pool sits there with its liquidity and its settings for 30 seconds before anyone can trade. Everyone watching knows the exact second it opens. A bot that reads the oracle can aim its buy at that second. A bot that does not read it sends buys that fail, and pays network fees for each failed try.</p><h3>How to read an Orca pool before your bot fires</h3><p>Everything a bot needs to price a trade on Orca is on chain. Here is the order to check it in.</p><ol><li><strong>Does the pool have an oracle?</strong> If not, the fee is fixed and the label is the fee. If yes, it is adaptive and the label is only the floor.</li><li><strong>What are the base fee and the ceiling?</strong> On a 1% adaptive pool, expect anything from 1% to 10%. On a 0.01% adaptive pool, the extra is usually smaller.</li><li><strong>Is there a trade enable time in the future?</strong> If yes, trading has not started. Time the buy to that second.</li><li><strong>Is there liquidity yet?</strong> A new pool with nothing in it is not a trade. Wait for your minimum.</li><li><strong>How high is the volatility right now?</strong> The oracle stores the accumulator. A high number means the next trade, yours included, pays more. This is the difference between seat one and seat two.</li><li><strong>What shape is the liquidity?</strong> A splash pool is full range and easy to read: the SOL in the pool is what you trade against. A concentrated pool can look deep at the current price and thin out a few percent higher, so a big buy moves the price much further than you expect.</li><li><strong>Is the coin itself safe to buy?</strong> The pool says nothing about the coin. Check creator holdings, top holders, mint and freeze authority, and any Token-2022 transfer fee.</li></ol><figure><img alt="Checklist of seven checks to read an Orca Whirlpool before a sniper buys: 1 does the pool have an oracle, 2 base fee and the 10% cap, 3 trade enable time, 4 is there liquidity yet, 5 current volatility, 6 splash pool or narrow concentrated range, 7 the coin itself: dev holdings, top 10 holders, mint and freeze authority and transfer fee. If all pass, buy now. If the pool is already running, skip it." src="https://cdn-images-1.medium.com/max/1024/1*Z6H9pTpSCSWPdr4RZgYhAQ.png" /></figure><h3>What each seat should do differently</h3><p>Now the useful part. Here is how to get into seat one more often, and what to do when you can’t.</p><p><strong>If you can land in the first seconds, buy.</strong> On an adaptive pool the open is the cheapest time to buy, which is not true on venues that tax speed. A slightly higher Jito tip and priority fee are worth it here. Pair them with a minimum liquidity filter so the bot buys the moment real liquidity arrives, not the moment the empty pool appears.</p><p><strong>If you missed the open, do not chase.</strong> The wave right after the open is the most expensive time to buy. If your bot sees a pool that is already moving fast, the fee is probably high and the price is already up. Skipping is often the right call. The FOMOEYES list shows that the fee can fall back to near base in a pause, but your bot cannot know when the next pause will come, so treat a pool that is already running as a pass.</p><p><strong>If the pool has a scheduled start, aim for the second.</strong> On a pool with a trade enable time, the start is known. Send the buy to land at that second, not before.</p><p><strong>Set exits for the worst case.</strong> If you buy at 1% and sell into the rush at 10%, a +30% target is not +30%. Start your first take-profit around +60% so a 10% exit fee still leaves a real gain. Add a trailing stop, which sells after the price has peaked and started to fall back. A falling price still moves ticks, so the fee does not disappear, but the trailing exit tends to fire after the fastest part of the move, when the accumulator has had time to start decaying.</p><p><strong>Use a timer.</strong> After about an hour, the fee usually drifts back toward base, and so does the interest in most new pools. If the trade has not worked by then, it often will not.</p><p><strong>Leave room for the fee in your slippage cap.</strong> The adaptive fee comes out of your trade. If your quote ignores it, a buy that pays 10% in fees looks like a buy with 10% slippage and fails. A 15% cap leaves room for a high-fee fill without also accepting a bad price on top of it.</p><h3>Orca sniper settings to start from</h3><p>These are starting points for the two triggers most people use on Orca. Change them to fit your size and how patient you are. They follow from everything above: wait for liquidity, check the coin, land early, and set exits high enough to survive a 10% fee each way.</p><p><strong>New pool sniper (any new Whirlpool):</strong></p><ul><li>Minimum liquidity: 10 SOL or more</li><li>Buy size: 0.1 to 0.3 SOL</li><li>Slippage cap: 15%</li><li>Priority fee: 0.001 SOL</li><li>Jito tip: 0.003 SOL</li><li>Dev holding cap: 10%</li><li>Top 10 holder cap: 30%</li><li>Maximum tax: 0%</li><li>Take-profit: +60% sell 50%, +150% sell 50%</li><li>Trailing stop: 20%</li><li>Stop-loss: 25%</li><li>Timer: 60 minutes</li></ul><p><strong>Creator watch sniper (pools made by wallets you choose):</strong></p><ul><li>Minimum liquidity: 5 SOL or more</li><li>Buy size: 0.2 to 0.5 SOL</li><li>Slippage cap: 15%</li><li>Priority fee: 0.001 SOL</li><li>Jito tip: 0.002 SOL</li><li>Dev holding cap: 15%</li><li>Top 10 holder cap: 40%</li><li>Maximum tax: 0%</li><li>Take-profit: +60% sell 50%, +200% sell 50%</li><li>Trailing stop: 25%</li><li>Stop-loss: 30%</li><li>Timer: 120 minutes</li></ul><figure><img alt="Table of starting Orca sniper settings. New pool: min liquidity 10 SOL, buy 0.1 to 0.3 SOL, slippage 15%, priority 0.001 and tip 0.003 SOL, dev cap 10%, top 10 cap 30%, max tax 0%, take-profit +60% and +150%, trailing 20%, stop-loss 25%, timer 60 min. Creator watch: 5 SOL, 0.2 to 0.5 SOL, 15%, tip 0.002, 15% and 40%, 0%, +60% and +200%, trailing 25%, stop-loss 30%, 120 min." src="https://cdn-images-1.medium.com/max/1024/1*-jT3dXk7dCdjIpVQb5c6Pg.png" /></figure><p>Why the creator watch version is looser: you already picked the wallet, so you trust its pools a little more, and you can afford a lower liquidity floor and a longer timer. Why the new pool version has a higher tip: with no creator to filter on, being early is your main edge, and on Orca being early is also cheaper.</p><p>If you want to go deeper on the settings themselves, we tested slippage, priority fee and tip values on 807 Pump.fun launches in <a href="https://medium.com/@dexsnipecom/best-sniper-bot-settings-for-pumpfun-slippage-priority-fee-and-tip-from-807-launches-76409e7a5040">Best Sniper Bot Settings for Pumpfun</a>, and checked which coin filters held up in <a href="https://medium.com/@dexsnipecom/pump-fun-sniper-bot-filters-we-checked-807-new-coins-to-see-which-rules-work-5c7a932f9e9e">Pump.fun Sniper Bot Filters</a>. The filters carry over to Orca almost as they are, because the coins are often the same coins.</p><h3>When things go wrong, and what to change</h3><ul><li><strong>Buys fill for far fewer tokens than the quote.</strong> The adaptive fee was high when you filled. Buy earlier, or skip pools that are already moving.</li><li><strong>Take-profits hit but the trade comes out flat.</strong> You sold into a high-fee wave. Raise your first target and add a trailing stop.</li><li><strong>The sniper fires but nothing fills.</strong> The pool was still empty or not yet tradable. Raise the minimum liquidity and check for a trade enable time.</li><li><strong>A big buy moves the price much more than expected.</strong> The pool has a narrow concentrated range. Lower the buy size or prefer splash pools.</li><li><strong>Every pool from one creator goes nowhere.</strong> Add that wallet to your skip list and move on.</li></ul><h3>Orca next to the other venues</h3><p>It helps to see Orca against the venues we have covered before, because each one rewards a different kind of timing.</p><ul><li><strong>Meteora:</strong> many launches open with an anti-sniper fee that starts high and falls over the first minute. There, waiting a few seconds can be much cheaper than being first. We broke that down second by second in <a href="https://medium.com/@dexsnipecom/meteora-sniper-bot-when-to-buy-a-launch-second-by-second-without-paying-the-anti-sniper-fee-5d0fea77d994">Meteora Sniper Bot: When to Buy a Launch</a>.</li><li><strong>PumpSwap:</strong> the pools that Pump.fun coins migrate into. The fee is flat, but the migrations come in three very different kinds, and only one of them is a real market. See <a href="https://medium.com/@dexsnipecom/pumpswap-sniper-bot-10-mistakes-that-turn-a-pump-fun-migration-snipe-into-a-loss-871f672d0819">PumpSwap Sniper Bot: 10 Mistakes</a>.</li><li><strong>Orca:</strong> the fee follows the price, not the clock. The first buyer pays least, the crowd pays most, and exits into a rush pay up to 10%.</li></ul><p>So there is no single rule for “how fast should my sniper be.” On Meteora’s high-fee launches, slow down. On Orca’s adaptive pools, be first or skip. On flat-fee pools, the usual race applies. A bot that runs the same timing on every venue will be wrong on at least one of them.</p><h3>What to look for in an Orca sniper bot</h3><ul><li><strong>Does it read Whirlpools directly?</strong> A bot that only watches launchpads will miss the pools created for coins after they leave the curve.</li><li><strong>Does it wait for liquidity?</strong> New Whirlpools start empty. A bot without a minimum liquidity filter fires at nothing.</li><li><strong>Does it check the coin, not just the pool?</strong> Dev holdings, top holders, mint and freeze authority, and transfer fees.</li><li><strong>Does it have a trailing stop?</strong> On a pool whose fee rises with the move, exiting after the peak matters.</li><li><strong>Do exits run when you are offline?</strong> A sniper that only sells while your browser tab is open will miss the exit that counts.</li><li><strong>How is it priced?</strong> On a venue where the pool fee can already reach 10%, a bot that adds its own percentage to every trade makes it worse.</li></ul><h3>Setting up an Orca sniper in DexSnipe</h3><p>DexSnipe supports Orca Whirlpools with three triggers: New pool, Creator watch and Copy trade. A creator watch or copy trade sniper can cover Orca together with the other venues DexSnipe supports, 12 in total, including Pump.fun, PumpSwap, LetsBonk, Raydium and Meteora. Every sniper has a minimum liquidity filter, a dev holding cap, a top 10 holder cap, mint and freeze checks, a maximum tax check, and exits with up to four take-profit steps, a stop-loss, a trailing stop, a timer and a dev-sell exit. Exits run on the server after you close the tab.</p><ol><li><strong>Open the panel and click New sniper.</strong> Choose New pool or Creator watch.</li><li><strong>Tick Orca,</strong> plus any other venue that supports the same trigger.</li><li><strong>Set the minimum liquidity</strong> so the sniper ignores empty pools.</li><li><strong>Set the buy:</strong> size, a 15% slippage cap, priority fee and tip.</li><li><strong>Set the filters:</strong> dev holding cap, top 10 cap and maximum tax.</li><li><strong>Set the exits:</strong> take-profits starting around +60%, a trailing stop, a stop-loss and a timer.</li><li><strong>Connect your wallet and save.</strong> Snipers are stored on your wallet address.</li><li><strong>Pick a plan, add a sniper wallet and arm it.</strong> Plans are 1 SOL for 7 days, 3 SOL for 30 days or 5 SOL for 90 days, with no fee per trade. Add the key of a separate funded wallet for the sniper to trade from, arm it, and close the tab.</li></ol><p>Use a separate wallet for sniping, and only fund it with what you are ready to lose. Most new coins lose value, and on adaptive pools the fee can take a large share of a fast trade.</p><h3>What this data does not tell you</h3><ul><li><strong>The fee findings come from a sample.</strong> 76 adaptive SOL pools created in 2026, first 100 transactions each. Older pools, non-SOL pairs and later trading may look different.</li><li><strong>Effective fees are fees divided by the amount in.</strong> A handful of trades with Token-2022 transfer fees or unusual routes gave odd values. We left out the ten that came out above 10%.</li><li><strong>The profit math is an illustration.</strong> It uses real fees from the FOMOEYES pool and an example 40% move. Real trades also pay network fees, tips and price impact.</li><li><strong>Coin suffixes are a shortcut.</strong> A mint ending in “pump” is almost always a Pump.fun coin, but we did not check each one.</li><li><strong>Wallets are not people.</strong> A wallet that creates many pools may be one person, a service or a market maker.</li></ul><h3>FAQ</h3><p><strong>What is an Orca sniper bot?</strong><br>A bot that watches Orca’s Whirlpools program for new pools, buys the moment one opens with real liquidity, and sells on rules you set. It can also follow chosen wallets and buy only the pools they create.</p><p><strong>What is an Orca adaptive fee pool?</strong><br>A Whirlpool whose fee has two parts: a fixed base fee and an adaptive part that rises when the price moves fast. Each one has an oracle account that stores the settings and the pool’s current volatility. 25,674 of the 167,036 Whirlpools we read were adaptive.</p><p><strong>How high can the Orca fee go?</strong><br>The program caps the total fee at 10%. On 22,468 of the 25,674 adaptive pools, the settings are steep enough to reach that cap. 34 of the 76 new pools we measured charged close to 10% on at least one trade in their opening.</p><p><strong>Do snipers pay the highest fee on Orca?</strong><br>Usually not the very first buyer. The first trades in a calm pool pay close to the base fee. In our sample, buys in the first ten seconds paid a median of 2.0%, and buys between one and five minutes paid a median of 5.05%. The crowd that follows pays the most.</p><p><strong>Does the adaptive fee apply when I sell?</strong><br>Yes. It applies to both buys and sells. Selling into a fast move is the most expensive time to sell, which is why take-profits on Orca should be set higher than on fixed-fee pools.</p><p><strong>What is an Orca splash pool?</strong><br>A full-range Whirlpool with a 1% fee tier and a tick spacing of 32896, which Orca recommends for token launches. 15,639 Whirlpools are splash pools, and 8,367 of them use adaptive fees.</p><p><strong>Can I see when trading starts on an Orca pool?</strong><br>On some. An adaptive pool’s oracle can hold a trade enable time, and the program refuses swaps before it. 85 pools have one, most of them Wavebreak graduations, and on every one trading opened exactly 30 seconds after the pool was created.</p><p><strong>Why did my Orca sniper fire but not buy anything?</strong><br>New Whirlpools are created empty and funded later. In 80 new adaptive SOL pools, none had liquidity in the creation transaction, and the median first swap came about a minute later. Set a minimum liquidity filter.</p><p><strong>Does DexSnipe take a fee on Orca trades?</strong><br>No. DexSnipe is a flat plan: 1 SOL for 7 days, 3 SOL for 30 days or 5 SOL for 90 days. You pay the network fee, your priority fee and tip, and the pool’s own trading fee, which on adaptive pools can change from one trade to the next.</p><h3>How we collected the data</h3><p>On November 6, 2026 we listed every Whirlpool account and every oracle account of Orca’s Whirlpools program. From each pool we read the config, tick spacing, fee tier, liquidity and token mints. From each oracle we read the pool it belongs to, the trade enable time, the adaptive fee settings and the current volatility. The oracle layout and the fee formula come from Orca’s open-source program on GitHub, and the maximum adaptive fee for each pool was computed from its settings and capped at the program’s 10% limit.</p><p>For fees, we read the first 100 transactions of 76 adaptive SOL pools created at even intervals through 2026, and decoded the trade event of every successful swap: amount in, LP fee and protocol fee. Buy or sell was decided from the swap direction and which side of the pool is SOL. The FOMOEYES list is one of those pools. For pool creation, we read the first 12 transactions of 80 adaptive SOL pools from 2026. Trade enable times were checked against each of the 85 pools’ first transaction. Pool dates come from the first transaction of 500 adaptive and 500 fixed-fee pools chosen at even intervals.</p><h3>Sources and further reading</h3><ul><li>Orca docs, adaptive fee pools (base fee, adaptive fee, and how price movement raises it): <a href="https://docs.orca.so/liquidity/concepts/adaptive-fees">docs.orca.so</a></li><li>Orca docs, understanding Whirlpool fees (how swap fees are calculated and split): <a href="https://docs.orca.so/developers/architecture/whirlpool-fees">docs.orca.so</a></li><li>Orca docs, ticks, tick spacing and fee tiers: <a href="https://docs.orca.so/liquidity/concepts/ticks-and-fees">docs.orca.so</a></li><li>Orca docs, trading fees (how fees are charged and shared): <a href="https://docs.orca.so/liquidity/concepts/trading-fees">docs.orca.so</a></li><li>Orca developer docs index (fee tiers by tick spacing, splash pools for token launches): <a href="https://docs.orca.so/llms.txt">docs.orca.so</a></li><li>Whirlpools fee rate manager on GitHub (the adaptive fee formula and the 10% limit): <a href="https://github.com/orca-so/whirlpools/blob/main/programs/whirlpool/src/manager/fee_rate_manager.rs">github.com/orca-so</a></li><li>Whirlpools oracle account on GitHub (oracle layout, trade enable time and its 72 hour maximum): <a href="https://github.com/orca-so/whirlpools/blob/main/programs/whirlpool/src/state/oracle.rs">github.com/orca-so</a></li><li>Whirlpools program on Solscan: <a href="https://solscan.io/account/whirLbMiicVdio4qvUfM5KAg6Ct8VwpYzGff3uctyCc">solscan.io</a></li><li>The FOMOEYES pool on Solscan: <a href="https://solscan.io/account/D21zjMevzPgC4dcd9J1MCXwER5ijnvheV1goiNvhGCm">solscan.io</a></li><li>SolanaFloor, Orca joins the launchpad wars with Wavebreak: <a href="https://solanafloor.com/news/orca-joins-launchpad-wars-with-wavebreak-what-different">solanafloor.com</a></li><li>Jito docs, tips and bundles on Solana: <a href="https://docs.jito.wtf/lowlatencytxnsend/">docs.jito.wtf</a></li></ul><p>Docs were checked in November 2026. Nothing here is financial advice. Meme coins are very risky, and most new tokens go to zero. Only trade with money you can afford to lose.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=dae478d23913" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Meteora Sniper Bot: When to Buy a Launch, Second by Second, Without Paying the Anti-Sniper Fee]]></title>
            <link>https://medium.com/@dexsnipecom/meteora-sniper-bot-when-to-buy-a-launch-second-by-second-without-paying-the-anti-sniper-fee-5d0fea77d994?source=rss-2654b58e3d1e------2</link>
            <guid isPermaLink="false">https://medium.com/p/5d0fea77d994</guid>
            <category><![CDATA[trading-bot]]></category>
            <category><![CDATA[memecoins]]></category>
            <category><![CDATA[crypto-trading]]></category>
            <category><![CDATA[solana-blockchain]]></category>
            <category><![CDATA[cryptocurrency]]></category>
            <dc:creator><![CDATA[DexSnipe]]></dc:creator>
            <pubDate>Sat, 14 Nov 2026 06:50:13 GMT</pubDate>
            <atom:updated>2026-11-14T06:50:13.748Z</atom:updated>
            <content:encoded><![CDATA[<h3>On Meteora the first second of a launch can be the most expensive moment to buy. Here is what a sniper pays at each point of the first five minutes, how to read the fee clock before your bot fires, and the settings that follow from it.</h3><p>Every sniper guide on Solana says the same thing: be first. On Meteora that advice can cost you most of your buy. Many launches on Meteora’s Dynamic Bonding Curve open with a trading fee of 50% to 99% and lower it over the next minute or so. Meteora calls it an anti-sniper fee, and it does exactly what the name says. A bot that fires in the first block of one of these launches pays the fee on the way in, often again on the way out, and calls the result bad luck. The good news is that the fee is not hidden. It sits in the launch config, it follows a formula, and a bot can work out the fee for any second before it sends a buy. This guide walks through the first five minutes of a Meteora launch, second by second, then shows how to read the clock and set up a Meteora sniper bot around it.</p><figure><a href="https://www.dexsnipe.com/"><img alt="DexSnipe banner: Snipe Meteora on your own rules. Launch, migration and new pool snipers with dev, holder and liquidity checks, exits run on the server. Button: Open dexsnipe.com, address https://www.dexsnipe.com/. Illustrated My snipers panel: Meteora DBC fresh launch with 10% dev cap, DAMM v2 migration with 20 SOL min liquidity, new pool with top 10 holders max 35%, copy trade off. Exits: take-profit steps, stop loss, trailing stop, timer, dev sell." src="https://cdn-images-1.medium.com/max/1024/1*hnaQS_iZcvSmP9YM7bcW2g.png" /></a><figcaption>DexSnipe is our own Solana sniper bot service.</figcaption></figure><h3>The short version</h3><ul><li><strong>More than a third of Meteora launches open with a fee of 50% or more.</strong> Of 9,217 recent launches on Meteora’s Dynamic Bonding Curve (DBC) with a meaningful graduation target, 3,435 started at 50% or higher and then stepped down.</li><li><strong>Seconds 1 and 2 are the worst time to buy those launches.</strong> Buyers in that window paid a median fee of 85%. Counting the fee, they paid about 7.3 times what the first buyer paid per token before fees.</li><li><strong>Half a minute later the picture flips.</strong> Most high-fee launches were back to a normal 2 to 3% fee within 30 to 60 seconds, and buyers in seconds 31 to 60 paid about 1.4 times the first buyer’s price, all in.</li><li><strong>On flat-fee launches the old rule still holds.</strong> More than half of Meteora launches charge a flat 1 to 5%. There, the first seconds were the cheapest time to buy and waiting cost more.</li><li><strong>So the bot has to know which launch it is looking at.</strong> A Meteora sniper that reads the fee schedule can buy flat-fee launches fast and wait out the high-fee ones. A bot that does not read it is guessing.</li></ul><h3>What a Meteora sniper bot actually buys</h3><p>Meteora is a set of Solana programs, and three of them matter for snipers.</p><ul><li><strong>The Dynamic Bonding Curve (DBC).</strong> This is Meteora’s launch program. Many launchpads build on it under their own brand, including Jupiter Studio, Believe and Bags. A token starts on a curve, the price rises as people buy, and when the curve reaches its target the token moves to a pool.</li><li><strong>DAMM v2.</strong> Meteora’s constant product pool. DBC tokens graduate into DAMM v2 pools, and anyone can also open a DAMM v2 pool by hand for any token.</li><li><strong>DLMM.</strong> Meteora’s concentrated liquidity pool. It is used more for established pairs than for brand new tokens.</li></ul><p>A Meteora sniper bot watches one or more of these and buys at a moment you pick: a fresh launch on the DBC, the DAMM v2 pool a launch graduates into, or a new DAMM v2 pool that someone opens. Then it sells on rules you set.</p><p>What makes Meteora different from Pump.fun or LetsBonk is that each launchpad on DBC chooses its own fee design. Meteora’s docs are open about it: the fee can start high so that bots racing for the first block do not take all the early supply, and it is capped at 99%. The fee is not burned either. Per Meteora’s docs, 20% of each DBC trading fee goes to the protocol and the rest is split between the launchpad partner and, where set up, the token creator. When your bot pays 85% on a buy, somebody collects it.</p><p>That changes the question a sniper has to answer. On Pump.fun the question is how fast you can land. On a Meteora launch the question is when the fee falls below the level where buying makes sense.</p><h3>Second 0: the buy inside the launch transaction</h3><p>Before anyone outside can trade, there is one buy that often goes through at a normal fee. Meteora’s launch configuration has an option called first swap with minimum fee. It lets the transaction that creates the pool include a single buy at the minimum fee, as long as it targets that new pool and carries no referral.</p><p>In practice, that buy usually belongs to the creator or the launchpad. In our sample of 137 high-fee launches, 120 buys landed inside the launch transaction itself. Their median fee was 2.0%, and they lost 2.2% of the SOL they spent to fees. The median size of those buys was 0.5 SOL.</p><p>This is not a trick. Many launchpads use the initial buy so creators hold some of their own token, and Meteora documents the option. It does change how you should read the first minute. On a launch built to tax snipers, the creator can buy first at a normal fee. That wallet is in profit the moment anyone pays 50% to get in after it. On these launches a dev holding cap and a dev-sell exit matter more, not less.</p><p><strong>What a sniper can do here:</strong> nothing. You cannot buy inside someone else’s launch transaction. What you can do is notice who did.</p><h3>Seconds 1 and 2: the most expensive moment on Meteora</h3><p>This is the window every classic sniper is built for, and on high-fee launches it is the worst possible time to buy.</p><p>In our 137 high-fee launches, 22 buys landed in seconds 1 and 2. They paid a median fee of 85.0%. Together they spent 24.7 SOL and lost 77.5% of it to fees before the price moved at all. Measured all in, fee included, the median buyer in this window paid 7.28 times the price the first outside buyer paid before fees.</p><p>Those buyers were not one careless trader. Across our sample, 141 buys paid a fee of 50% or more. They came from 101 different wallets, spent 78.2 SOL and handed 60.9 SOL of it to the fee. That is what you would expect from bots set to fire on any new launch with no idea what the fee was.</p><p><strong>What a sniper should do here:</strong> on a launch whose fee starts at 50% or more, nothing. A big Jito tip to land in this window is paying twice to lose.</p><h3>Seconds 3 to 5: still a coin flip against the fee</h3><p>Three seconds in, most schedules have stepped down a few times. They are still steep. Buyers in seconds 3 to 5 paid a median fee of 50.1% and lost 55.6% of what they spent to fees. All in, they paid 2.20 times the first buyer’s pre-fee price.</p><p>The price itself barely moved in these seconds, and that is worth noticing. Very few people buy into a 50% fee, so the curve does not climb much while the fee is high. The fee protects the price as much as it punishes the buyer. That is why waiting a little longer does not mean losing the entry. On a high-fee launch the entry is still there; only the toll is falling.</p><figure><img alt="Line chart of two Meteora DBC fee clocks over the first 120 seconds. The sprint starts at 96% and steps down every second for 50 steps, used by 216 launches: 44% at 10 s, 9.4% at 30 s, under a 5% limit at about second 39, 2% from second 50. The squeeze starts at 51.9% and steps down every 5 seconds for 120 steps, used by 811 launches: 43.5% at 30 s, 36.4% at 60 s, still 25.5% at 2 minutes." src="https://cdn-images-1.medium.com/max/1024/1*1BvaK6ONurlqMrRaYMz_kg.png" /></figure><h3>Seconds 6 to 10: the median looks fine, the average does not</h3><p>This is where many bots that “wait a few seconds” end up, and the numbers are tricky. The median buyer between 6 and 10 seconds paid a fee of only 9.5%. But the buyers who landed early in that window paid much more, so 28.3% of all SOL spent in those five seconds went to fees. All in, buyers paid 1.75 times the first buyer’s price.</p><p>In other words, “wait 5 seconds” is not a rule. On a fast schedule it lands you near the end of the high fee. On a slow one it lands you in the middle of it. The only reliable rule is to compute the fee for the second you plan to buy, which we come back to below.</p><h3>Seconds 11 to 30: the busy window</h3><p>Once the fee is down to a few percent, buying picks up and the price starts to climb. Between 11 and 30 seconds we counted 384 buys, the busiest window in our sample. Their median fee was 2.5%, they lost 15.0% of their SOL to fees overall, and they paid 1.55 times the first buyer’s price.</p><p>The 15% loss is higher than the median fee because some schedules are still running at second 15 or 20. A fast schedule that starts at 96% and drops one step per second is at about 44% at second 10 and 20% at second 20. A buyer at second 12 on that launch still pays a lot.</p><h3>Seconds 31 to 60: where most high-fee launches become normal</h3><p>By half a minute in, the typical high-fee schedule has run out. Across every high-start schedule we found on the DBC program, the median time from opening to the final fee was 48 seconds. A quarter finished in under 15 seconds and three quarters within a minute. The final fee was usually around 2%.</p><p>Buyers in seconds 31 to 60 paid a median fee of 2.8% and lost 6.0% of their SOL to fees overall. All in, they paid 1.38 times the first buyer’s pre-fee price. Compare that with 7.28 times for the buyer at second 2. On a high-fee launch, being half a minute late was more than five times cheaper per token than being first in line.</p><h3>Minutes 1 to 5, and after</h3><p>From the first minute on, the fee is at its final level on almost every launch, and the cost per token depends on the price alone. Buyers between 1 and 5 minutes paid a median fee of 2.0% and 1.30 times the first buyer’s price. Buyers after 5 minutes paid 1.08 times.</p><p>That last number surprises people. On many high-fee launches the price five minutes in was close to where the first outside buyer got in, because the early fee kept the curve from running. Not every launch is like that, and some do run hard once the fee drops. But the idea that a Meteora launch is “gone” after the first block does not hold up for the high-fee kind.</p><h3>The same clock on a flat-fee launch is the opposite</h3><p>Now take the 55% of Meteora launches that charge a flat fee of 1 to 5% from the first second. We ran the same count on 93 of them. The fee barely changed with time: a median of 1.0% to 1.3% in every window, and 1.1% to 1.8% of SOL lost to fees. What changed was the price.</p><ul><li><strong>Seconds 1 to 2:</strong> 1.20 times the first buyer’s price.</li><li><strong>Seconds 3 to 5:</strong> 1.36 times.</li><li><strong>Seconds 6 to 30:</strong> 1.44 times.</li><li><strong>Seconds 31 to 60:</strong> 1.91 times.</li><li><strong>After 5 minutes:</strong> 1.46 times.</li></ul><p>On a flat-fee launch, every buyer pushes the price up for the next one, so speed pays here just like on Pump.fun. This is why a Meteora sniper needs two modes. The same “wait 30 seconds” rule that saves you on a high-fee launch costs you on a flat one.</p><figure><img alt="Bar chart of all-in cost per token vs the first outside buyer’s pre-fee price, by when the buy landed. 137 high-fee launches: 7.28x at 1–2 s, 2.20x at 3–5 s, 1.75x at 6–10 s, 1.55x at 11–30 s, 1.38x at 31–60 s, 1.30x at 1–5 min, 1.08x after 5 min. 93 flat-fee launches: 1.20x, 1.36x, 1.44x, 1.44x, 1.91x, 1.58x, 1.46x. Waiting was cheaper on high-fee launches, being early was cheaper on flat-fee ones." src="https://cdn-images-1.medium.com/max/1024/1*IbV2FvJNSjAHtv25fZyBug.png" /></figure><h3>How a sniper bot reads the fee clock before it buys</h3><p>Every DBC launch config stores four numbers for its base fee:</p><ol><li><strong>Cliff fee.</strong> The fee when trading opens.</li><li><strong>Number of periods.</strong> How many times the fee steps down.</li><li><strong>Period frequency.</strong> How long each step lasts, in seconds or in slots.</li><li><strong>Reduction factor.</strong> How much each step removes.</li></ol><p>If all three schedule numbers are zero, the launch charges the cliff fee forever. That is how a flat fee is stored. Otherwise the schedule has one of two shapes, both in Meteora’s docs:</p><ul><li><strong>Linear:</strong> fee = cliff fee minus (periods passed times reduction). The fee drops by the same amount each step.</li><li><strong>Exponential:</strong> fee = cliff fee times (1 minus reduction) to the power of periods passed. The fee drops fast at first and then slows down.</li></ul><p>The clock starts at the pool’s activation point, not when the token was created, and it runs on time, not on trades. Two seconds after activation the fee is the same whether one person has traded or a thousand have. That is what makes it predictable. A bot that knows the four numbers and the activation time can compute the fee at any second, and can compute the first second at which the fee drops under a limit you set.</p><p><strong>A worked example.</strong> Take the most used fast schedule we found among recent launches: a 96% start that steps down every second for 50 steps. Rebuilt with Meteora’s formula, its fee was:</p><ul><li>96.0% at second 0</li><li>82.2% at second 2</li><li>65.2% at second 5</li><li>44.3% at second 10</li><li>20.4% at second 20</li><li>9.4% at second 30</li><li>2.0% from second 50 on</li></ul><p>If your limit is “pay no more than 5% in fees,” the bot should wait until about second 39 on this launch. If your limit is 3%, about second 45. On a flat 2% launch, the same limit means “buy now.”</p><h3>Two clocks that look the same in a launch feed</h3><p>Here is why a fixed delay does not work. The second most used high-fee schedule in our month was set by a single launchpad on 811 launches. It started at 51.9% and stepped down every 5 seconds for 120 steps:</p><ul><li>51.9% at second 0</li><li>48.9% at second 10</li><li>43.5% at second 30</li><li>36.4% at second 60</li><li>25.5% at second 120</li></ul><p>It never goes as high as the 96% sprint, but it is still taking a quarter of every buy two minutes in, and it runs for ten minutes. A bot tuned to “wait 40 seconds” would buy the sprint at a fair fee and buy this one at about 40%. Both launches look the same in a new launch feed: a new curve, a name, a ticker. Only the config tells them apart.</p><p><strong>Slots or seconds.</strong> A config can measure its clock in slots instead of seconds. Solana slots got shorter in 2026, and slot timing drifts under load. For time-based configs the formula is exact. For slot-based ones, treat the result as an estimate and leave a second of margin. All the 30 day figures in this guide come from time-based launches, which are the large majority.</p><p><strong>The fee on top of the fee.</strong> 913 of the 9,217 launches also switched on Meteora’s dynamic fee. It rises with volatility and decays when trading calms down. On a launch, volatility is highest exactly when snipers arrive, so the dynamic part adds the most where the scheduled part is already high. It cannot be computed from the config alone. A bot has to read the pool’s volatility state when it quotes.</p><p><strong>Token fees are separate.</strong> 706 recent launches used Token-2022 mints, and 26 set a transfer fee on the token itself. That fee is charged by the token program on every transfer, including your buy and your sell, on top of the pool fee. DBC also supports transfer hook launches, where a custom program runs on every transfer during the curve. Meteora removes the hook when the curve completes. Until then, a bot should read it before buying.</p><h3>Pools whose clock never moves: the 70% fee trap</h3><p>The anti-sniper fee on launches is temporary by design. There is another high fee on Meteora that is not. DAMM v2 lets anyone open a pool and pick its base fee. When we decoded the 88,542 DAMM v2 pools opened over 30 days, 4,366 had a fixed fee of 50% or more and no schedule to lower it. 2,709 of them were set to exactly 70%.</p><p>Most of them paired SOL with a token that already trades on Pump.fun or PumpSwap. 2,947 of the SOL pools were for mints ending in “pump”. 37 wallets opened all the SOL pools, and the ten busiest opened 3,762 of them. A new pool for an existing token is exactly the event a “new pool” sniper or a copy bot reacts to. The pools’ own fee counters recorded roughly 1,116 SOL in fees over the month, most of it paid by trades routed through third-party trading programs. Whoever routed them did not check the fee.</p><p>When we read the pools, only 477 of the 4,366 still had any liquidity, and almost all of the SOL pools held less than 1 SOL. The pattern fits pools that are opened, collect fees from whatever buys into them, and are then emptied.</p><p><strong>How a bot spots one before buying:</strong></p><ul><li><strong>The base fee is far above what a real market needs.</strong> A pool a DBC launch graduates into can have a custom fee of at most 10%, and busy pools charge far less.</li><li><strong>There is no scheduler.</strong> A high fee that is not set to fall is not protection. It is the point of the pool.</li><li><strong>The token already has a busy market elsewhere.</strong> A brand new launch has one pool. A trap pool is a second or third pool for a known token.</li><li><strong>The opener has opened hundreds of these.</strong></li></ul><p>The fix is a single setting: a maximum pool fee on every “new pool” sniper. 3% is enough for any normal market.</p><h3>After the curve: the migration pool has its own fee and its own lock</h3><p>When a DBC curve fills, the token moves into a DAMM v2 pool. That is where migration snipers buy. The 90% opening fees of the curve do not follow the token there; a custom fee on a graduated pool must sit between 0.1% and 10%. Two other settings in the launch config do matter, and both can be read before the pool exists.</p><p><strong>The market cap scheduler.</strong> A graduated pool can use a fee that only steps down as the price rises above the opening price, with an expiry after which it settles at its final level. If the price goes nowhere, the fee stays high until the expiry. In our 30 days, 479 new DAMM v2 pools used one. For a migration sniper that is an awkward deal: you pay the higher fee exactly when the token is not doing well, and again if you sell into weakness.</p><p><strong>How much liquidity stays locked.</strong> Meteora requires at least 10% of migrated liquidity to stay permanently locked at day one. The config decides the other 90%: permanently locked, vesting over time, or in positions the partner and creator can withdraw. In 8,625 of the 9,217 recent launches, none of the graduated liquidity was set to be freely withdrawable. In 564 launches, 76 to 90% of it was. Those are the launches where a migration snipe can be followed by a large withdrawal, the same pattern we saw on hand-opened Raydium pools in <a href="https://medium.com/@dexsnipecom/raydium-sniper-bot-on-github-9-checks-before-you-run-one-2566b90171e1">our guide to checking Raydium sniper bots on GitHub</a>.</p><p>If you snipe migrations on Pump.fun as well, the trade there looks quite different, with no anti-sniper fee but a lot of instant and empty pools. We went through it in <a href="https://medium.com/@dexsnipecom/pumpswap-sniper-bot-10-mistakes-that-turn-a-pump-fun-migration-snipe-into-a-loss-871f672d0819">10 mistakes PumpSwap migration snipers make</a>.</p><h3>Alpha Vault: the buyer who comes before second 0</h3><p>If being first is what you are after, Meteora has an answer that does not involve a bot. An Alpha Vault is attached to a launch pool and collects deposits before trading opens. During a protected window before public activation, the vault buys from the pool for all depositors at once. Public trading, and every sniper, starts after that.</p><p>Vaults run in two modes. Pro rata accepts more deposits than it can fill, gives everyone a proportional share and refunds the rest. First come first served fills deposits until a cap is reached. Tokens bought by the vault are released linearly between a start and an end point, so vault buyers cannot all sell in the first block.</p><p>For a sniper this matters twice. On a launch with an Alpha Vault you are never the first buyer, however fast you are; the vault is. And the vault’s tokens are released on a schedule, so the selling pressure from them is predictable. If a project you follow announces a vault, the deposit is the early entry, not the snipe.</p><figure><img alt="Diagram of how a Meteora sniper reads the fee clock. Four numbers in the DBC launch config: cliff fee, number of periods, period frequency in seconds or slots, reduction factor. All three schedule numbers at zero means a flat fee. Linear: fee = cliff fee minus periods passed times reduction. Exponential: fee = cliff fee times (1 minus reduction) to the periods passed. If the fee is under your limit, buy now; if not, compute the first second it drops under and buy then. Add dynamic and transfer f" src="https://cdn-images-1.medium.com/max/1024/1*GHwLVggitOu7KuXbcBmrCg.png" /></figure><h3>Meteora sniper bot settings, by launch type</h3><p>Most sniper settings carry over from other Solana venues. Meteora needs one extra check, the fee, and a different setup for each kind of launch. Where a bot has no fee limit, the trigger choice does the same job. These are starting points, not promises.</p><p><strong>Flat-fee DBC launches (fee 1 to 5% from the start)</strong></p><ul><li>Trigger: fresh launch.</li><li>Fee worth paying: up to 3%.</li><li>Buy size: 0.1 to 0.25 SOL. Slippage cap: 15 to 20%.</li><li>Priority fee: 0.001 SOL. Jito tip: 0.002 SOL. Here the first block really is the cheapest, so a normal tip is worth it.</li><li>Dev holding cap: 10%.</li><li>Exits: +100% sell 50%, +300% sell 50%, stop-loss 35%.</li></ul><p><strong>High-fee DBC launches (fee 50% or more, then falling)</strong></p><ul><li>Trigger: either skip the launch and snipe the migration, or buy only once the computed fee is under your limit.</li><li>Fee worth paying: up to 3 to 5%.</li><li>Buy size: 0.1 to 0.25 SOL. Slippage cap: 15 to 25%.</li><li>Priority fee: 0.0005 SOL. Jito tip: 0.001 SOL. Once the schedule has run down there is no first-slot race left to win.</li><li>Dev holding cap: 5 to 10%, because the creator often bought at the minimum fee.</li><li>Exits: +60% sell 50%, +200% sell 50%, stop-loss 25 to 30%.</li></ul><p><strong>New DAMM v2 pools</strong></p><ul><li>Trigger: new pool, only after reading its base fee and scheduler.</li><li>Fee worth paying: up to 2 to 3%. This one setting skips every 70% trap pool.</li><li>Buy size: 0.1 to 0.5 SOL. Slippage cap: 10 to 15%.</li><li>Priority fee: 0.0005 SOL. Jito tip: 0.001 SOL.</li><li>Minimum liquidity and LP burned or locked: on.</li><li>Exits: +40% sell 50%, +120% sell 50%, stop-loss 20 to 25%.</li></ul><p><strong>Slippage is not a fee check.</strong> A slippage cap protects you against the price moving between quote and fill. Whether it also protects you against the fee depends on how the bot builds its quote. If the quote ignores the pool fee, a 70% fee shows up as a fill far below the quote and the slippage cap stops the buy. If the quote includes the fee, the bot accepts it as the expected result. That is why the fee has to be checked on its own, before the buy. We covered how slippage, priority fee and tip work together in <a href="https://medium.com/@dexsnipecom/best-sniper-bot-settings-for-pumpfun-slippage-priority-fee-and-tip-from-807-launches-76409e7a5040">our sniper bot settings guide</a>.</p><h3>Exits on Meteora: the fee cuts both ways</h3><p>On the DBC, sells pay the scheduled fee too. A stop-loss that fires in the first seconds of a high-fee launch sells into the same 80% fee you avoided on the way in. A very short timer is a bad idea on these launches for the same reason. Give positions time to get past the schedule before any automatic sell can fire, unless you accept paying the fee twice.</p><p>Three more exit rules that fit Meteora:</p><ul><li><strong>Make the first take-profit return your cost.</strong> Sell half at a level that gets your SOL back. The rest can ride.</li><li><strong>Keep dev-sell on.</strong> On launches where the creator bought at the minimum fee, the creator selling is the clearest warning you will get.</li><li><strong>Run exits on a server.</strong> DBC curves and DAMM v2 pools do not hold limit orders. Every exit is your bot watching the price and sending a sell. If the bot only runs while a tab is open, the exit only runs while you are watching.</li></ul><h3>How to set up a Meteora sniper bot, step by step</h3><ol><li><strong>Make a separate sniper wallet.</strong> Create a new Solana wallet and fund it with only the SOL you plan to trade. Never give any bot a seed phrase.</li><li><strong>Decide which Meteora market you are sniping.</strong> Fresh launches, migrations and new pools behave so differently that each deserves its own sniper.</li><li><strong>Create the sniper and pick the trigger.</strong> Tick Meteora, plus any Meteora-based launchpad you want to include, such as Jupiter Studio, Believe or Bags. Treat each launchpad as its own venue so you can give it its own rules.</li><li><strong>Handle the fee.</strong> If your bot has a fee limit, set 3 to 5% for launches and 2 to 3% for new pools. Most bots do not have one. In that case, point the fresh launch sniper only at launchpads you know charge a flat fee, and catch the high-fee launches with a migration sniper instead.</li><li><strong>Set the buy.</strong> Size, slippage, priority fee and tip from the lists above. Lower the tip for high-fee launches.</li><li><strong>Set the filters.</strong> Mint and freeze authority off, a dev holding cap, a top holder cap, a bundled launch check, and a transfer tax cap of 0%.</li><li><strong>Build exits that survive the schedule.</strong> A first take-profit that returns your cost, a stop-loss, dev-sell on, and no very short timer on launches.</li><li><strong>Arm it small and read the first ten trades.</strong> For each one, check the fee you actually paid. If any buy paid more than your limit, the bot is not reading the clock.</li><li><strong>Change one thing at a time.</strong> Fee limit first, then size, then exits.</li></ol><figure><img alt="Meteora sniper settings by launch type. Flat-fee launches: fresh launch, fee limit 3%, 0.1 to 0.25 SOL, priority 0.001 and tip 0.002 SOL, dev cap 10%, take-profit +100% and +300%, stop-loss 35%. High-fee launches: wait or snipe the migration, fee limit 3 to 5%, priority 0.0005 and tip 0.001, dev cap 5 to 10%, +60% and +200%, stop 25 to 30%. New DAMM v2 pools: max pool fee 2 to 3%, 0.1 to 0.5 SOL, LP burned or locked, +40% and +120%, stop 20 to 25%." src="https://cdn-images-1.medium.com/max/1024/1*AuY3fAK3ql_HS09Kk8oGDQ.png" /></figure><h3>What to ask before you trust a Meteora sniper bot</h3><p>A bot built for flat-fee venues can be fast and still get all of these wrong.</p><ul><li><strong>Does it read the launch config’s fee schedule</strong> and compute the current fee from the activation point, including whether the clock runs in slots or seconds?</li><li><strong>Does it add the dynamic fee</strong> from the pool’s volatility state, or only the base fee?</li><li><strong>Does it check a DAMM v2 pool’s base fee</strong> before buying a new pool, and skip fixed high-fee pools?</li><li><strong>Does it read the token program,</strong> including transfer fees and transfer hooks on Token-2022 launches?</li><li><strong>Does it know the migration pool’s fee and liquidity lock</strong> before it fires a migration buy?</li><li><strong>Does it trade from a separate wallet you control,</strong> so a mistake costs only what is in that wallet?</li><li><strong>Does it take a share of every trade?</strong> On top of a Meteora fee, a 1% bot fee on each buy and sell adds up fast.</li></ul><p>There are open source DBC snipers on GitHub, many built for one launchpad during a hype cycle. Check when each was last updated. DBC has changed its migration options, closed its older rate limiter fee to new launches, and changed its swap event format more than once. A bot that decodes an old event will misread fees or miss trades. If you want to compare how Meteora launchpads differ from each other, <a href="https://medium.com/@dexsnipecom/solana-launchpads-compared-the-hidden-rules-of-letsbonk-jupiter-studio-bags-boop-believe-and-87cef7c5a86d">our comparison of the hidden rules of Solana launchpads</a> covers Jupiter Studio, Bags and Believe side by side.</p><h3>Questions people ask about Meteora sniper bots</h3><h4>What is a Meteora sniper bot?</h4><p>A bot that buys tokens on Meteora as soon as they become tradable: new launches on the Dynamic Bonding Curve, the DAMM v2 pool a launch graduates into, or a new DAMM v2 pool someone opens by hand. On Meteora a useful sniper also reads the pool’s fee before it buys.</p><h4>What is Meteora’s anti-sniper fee?</h4><p>A launch setting on the Dynamic Bonding Curve and DAMM v2. The pool opens with a high trading fee, called the cliff fee, and lowers it step by step on a schedule. Meteora caps the fee at 99%. In the launches we sampled, the fee usually fell from 50 to 99% to about 2% within a minute.</p><h4>When should I buy a Meteora launch?</h4><p>On a flat-fee launch, as early as you can, because the price rises with every buyer. On a launch with a high, falling fee, after the fee drops below your limit. On the most common fast schedule that was around 40 to 45 seconds after opening.</p><h4>How much do snipers lose to fees on Meteora launches?</h4><p>In our sample of 137 high-fee launches, buyers in the first two seconds paid a median fee of 85% and lost 77.5% of the SOL they spent. Across all buys in the sample, 109.2 SOL of 685.5 SOL went to fees.</p><h4>How do I check a Meteora pool’s fee before buying?</h4><p>Meteora’s pool page shows the base fee and whether a fee scheduler is on. For DBC launches the schedule sits in the launch config: starting fee, number of periods, period length and reduction. A sniper bot should read these from the chain and compute the current fee before it sends a buy.</p><h4>What are the 70% fee pools on Meteora?</h4><p>DAMM v2 pools opened with a fixed fee of 50 to 99% and no schedule to lower it, most of them for tokens that already trade elsewhere. In the 30 days we checked, 4,366 were opened, 2,709 at exactly 70%. A maximum fee setting of 3% on a new pool sniper skips all of them.</p><h4>Can I snipe a Meteora launch before trading opens?</h4><p>Not with a bot. If the launch has an Alpha Vault, depositing into the vault is the way to buy before public trading. Outside of that, the only buy before outside traders is the one inside the launch transaction, which usually belongs to the creator.</p><h4>Does the anti-sniper fee apply when I sell?</h4><p>Yes. On the Dynamic Bonding Curve, sells pay the scheduled fee too. That is why a stop-loss or timer that can fire in the first seconds of a high-fee launch is a bad idea.</p><h4>Is a Meteora sniper bot safe?</h4><p>The bot is only as safe as its wallet setup. Use a separate wallet with only trading money, never share a seed phrase, and read any open source bot before running it. The trade itself is high risk: most new tokens lose value.</p><h3>How we got these numbers</h3><p>We listed every account of Meteora’s Dynamic Bonding Curve and DAMM v2 programs through a Solana RPC node in November 2026 and decoded them with the layouts in Meteora’s published program IDL. For the DBC we read each launch config’s base fee mode, cliff fee, schedule, dynamic fee flag, activation type, quote mint, graduation target, token type and liquidity split, and joined them to every launch pool. The 30 day set covers time-based launches with a graduation target of at least 5 SOL or 500 USDC, which leaves out test launches and token farming. That gave 9,217 launches.</p><p>For the second-by-second numbers we drew a random sample of launches from the previous seven days: 137 with a starting fee of 50% or more on a falling schedule, and 93 with a flat fee of 2% or less, capped per launchpad so no single partner dominated. We read the first 45 successful transactions of each pool and decoded Meteora’s own swap events, which record the SOL put in, tokens received, fee taken and time. Fee share is the fee divided by the SOL put in. All-in cost per token is SOL put in divided by tokens received, relative to the pre-fee price of the first buy outside the launch transaction. For DAMM v2 we decoded pools activated in the last 30 days and read base fee, scheduler, liquidity, lock and fee counters, and cross-checked sampled pools against Meteora’s data API.</p><p>Seven days of swaps is a snapshot. Launchpads change their configs, and the mix of high-fee and flat-fee launches will move. The way to read a schedule, and the reason to read it, will not.</p><h3>Sources and further reading</h3><ul><li>Meteora docs, DBC fees (fee layers, the 99% cap, how the fee is split): <a href="https://docs.meteora.ag/core-products/dbc/fees/overview">docs.meteora.ag</a></li><li>Meteora docs, DBC fee scheduler (linear and exponential schedules and their formulas): <a href="https://docs.meteora.ag/core-products/dbc/fees/fee-scheduler">docs.meteora.ag</a></li><li>Meteora docs, DBC rate limiter (the older size-based fee mode): <a href="https://docs.meteora.ag/core-products/dbc/fees/rate-limiter">docs.meteora.ag</a></li><li>Meteora docs, DBC launch configuration (including first swap with minimum fee): <a href="https://docs.meteora.ag/core-products/dbc/launch-configurations">docs.meteora.ag</a></li><li>Meteora docs, DBC migration and liquidity (graduation to DAMM v2 and the 10% day one lock): <a href="https://docs.meteora.ag/core-products/dbc/migration-and-liquidity">docs.meteora.ag</a></li><li>Meteora docs, DAMM v2 market cap scheduler: <a href="https://docs.meteora.ag/core-products/damm-v2/fees/market-cap-scheduler">docs.meteora.ag</a></li><li>Meteora docs, what is Alpha Vault: <a href="https://docs.meteora.ag/helper-products/alpha-vault/what-is-alpha-vault">docs.meteora.ag</a></li><li>Meteora Dynamic Bonding Curve SDK and program IDL: <a href="https://github.com/MeteoraAg/dynamic-bonding-curve-sdk">github.com/MeteoraAg</a></li><li>Meteora Dynamic Bonding Curve program on Solscan: <a href="https://solscan.io/account/dbcij3LWUppWqq96dh6gJWwBifmcGfLSB5D4DuSMaqN">solscan.io</a></li><li>Meteora DAMM v2 program on Solscan: <a href="https://solscan.io/account/cpamdpZCGKUy5JxQXB4dcpGPiikHawvSWAd6mEn1sGG">solscan.io</a></li><li>Jito docs, tips and bundles on Solana: <a href="https://docs.jito.wtf/lowlatencytxnsend/">docs.jito.wtf</a></li></ul><p>Docs were checked in November 2026. Nothing here is financial advice. Meme coins are very risky, and most new tokens go to zero. Only trade with money you can afford to lose.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5d0fea77d994" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[PumpSwap Sniper Bot: 10 Mistakes That Turn a Pump.fun Migration Snipe Into a Loss]]></title>
            <link>https://medium.com/@dexsnipecom/pumpswap-sniper-bot-10-mistakes-that-turn-a-pump-fun-migration-snipe-into-a-loss-871f672d0819?source=rss-2654b58e3d1e------2</link>
            <guid isPermaLink="false">https://medium.com/p/871f672d0819</guid>
            <category><![CDATA[memecoins]]></category>
            <category><![CDATA[trading-bot]]></category>
            <category><![CDATA[solana-blockchain]]></category>
            <category><![CDATA[cryptocurrency]]></category>
            <category><![CDATA[crypto-trading]]></category>
            <dc:creator><![CDATA[DexSnipe]]></dc:creator>
            <pubDate>Fri, 13 Nov 2026 14:21:49 GMT</pubDate>
            <atom:updated>2026-11-13T14:21:49.268Z</atom:updated>
            <content:encoded><![CDATA[<h3>Sniping a coin as it moves from the Pump.fun curve to PumpSwap looks simple. Most of the money is lost in ten small, fixable mistakes. Here is each one, how to spot it, and the setting that fixes it.</h3><p>A PumpSwap sniper bot has one job. When a Pump.fun coin fills its bonding curve, the coin migrates to PumpSwap, Pump.fun’s own exchange, and a new pool opens. The bot buys that pool in its first blocks and sells on rules you set. On paper it is the cleanest trade in the Pump.fun world: the coin has already proved that people wanted it, and now it trades on a real pool with real depth. In practice, people who run migration snipers tend to lose money in the same few ways, over and over. None of them is about speed. They are about which pools the bot buys, how much it pays to get in, and how it gets out. This guide goes through the ten we see most, in the order they usually bite, and ends with a starter setup that avoids all of them.</p><figure><a href="https://www.dexsnipe.com/"><img alt="DexSnipe banner: Snipe only the migrations worth it. PumpSwap pools checked at migration for liquidity, dev share and top 10 holders, with exits running on the server. A button reads Open dexsnipe.com with the address https://www.dexsnipe.com/. An illustrated My snipers panel shows a Pump.fun to PumpSwap migration sniper with min 20 SOL liquidity, a holder check with a 10% dev cap and 35% top 10 cap, exits at +20% for half and +50% for the rest with a 5 minute timer, and a note that Mayhem pools" src="https://cdn-images-1.medium.com/max/1024/1*8NsWdIJNA4gP6Y-FV0V2_A.png" /></a><figcaption>DexSnipe is our own Solana sniper bot service.</figcaption></figure><h3>The short version</h3><ul><li><strong>Not every migration is a market.</strong> In 247 SOL-paired Pump.fun migrations we traced in early November 2026, about half graduated within five seconds of the coin being created, and nearly a third were Mayhem pools that opened with a fraction of a SOL. Only about one in five had a curve that filled through ordinary trading.</li><li><strong>Three filters do most of the work.</strong> A minimum liquidity, a dev holding cap and a top 10 holder cap skip the two kinds of migration you do not want, before your bot sends a buy.</li><li><strong>Being first is worth less than it sounds.</strong> Since October 2026, the buy that completes a Pump.fun curve can also buy into the pool before it opens. In half of the organic migrations we traced, someone had done that.</li><li><strong>Tips eat the edge.</strong> On organic pools, first-block buyers made a median 13% before costs and 7.4% after. The heaviest tippers gave back the most.</li><li><strong>It is a five-minute trade.</strong> The median first-block buyer sold after 130 seconds. A timer and close take-profit steps fit the trade better than a moonshot target.</li></ul><h3>How a Pump.fun migration to PumpSwap works</h3><p>Every Pump.fun coin starts life on a bonding curve. The curve sells tokens at a price that rises with each buy. When enough SOL has gone in, the curve is complete. From that moment, nobody can trade on the curve anymore, and the coin waits to migrate.</p><p>Migration means the Pump program moves the tokens left in the curve and the SOL the curve raised into a new pool on PumpSwap. PumpSwap launched in March 2025 so that graduated coins would stay inside Pump.fun instead of leaving for Raydium. The pool it creates is called the canonical pool. It is the one pool that the Pump program itself makes for that coin, and its creator field points at a Pump program address, which is how wallets and bots can tell it apart from any other pool for the same token.</p><p>Two details matter for a sniper. First, the migrate instruction takes the wallet that calls it as a signer, so anyone can open the pool once the curve is complete. Some wallets specialise in opening pools and buy in the same block. Second, a canonical pool charges a fee that depends on the coin’s market cap. Pump moved both its curves and its canonical pools to market cap fee tiers in September 2025, so the fee on your buy can change as the coin grows. Non-canonical pools for the same coin use a flat fee set separately.</p><p>A PumpSwap sniper bot listens for the migration, works out the canonical pool, runs its filters on that pool, and sends a buy. Everything after that is exits.</p><h3>Mistake 1: Firing on every migration</h3><p>The default setting on many bots is “buy every graduation.” It sounds reasonable. A graduation is a coin that filled its curve, so it should be one of the better coins of the day.</p><p>In late 2026 that assumption is wrong more often than right, because three very different things now end in a migration:</p><ul><li><strong>Instant migrations.</strong> A wallet creates a coin and buys the whole curve in the same transaction. The pool opens seconds later, often opened by the creator. In our sample, 121 of 247 migrations (49%) opened within five seconds of the coin’s creation, and 119 different creators were behind those 121 coins. This is a common technique, not one operator.</li><li><strong>Mayhem migrations.</strong> Mayhem Mode is a Pump.fun option that gives a coin an automated trading agent for its first day. 75 of our 247 migrations (30%) were Mayhem coins. Their pools opened almost empty.</li><li><strong>Organic migrations.</strong> People traded the coin on the curve and the curve filled. 51 of 247 (21%). The median one took 6.7 minutes from creation to migration and had 117 curve transactions on the way.</li></ul><p>In a migration feed, these look the same: a curve completes, a pool opens, an event fires. Once the pool trades, they behave nothing alike. A bot that buys all of them is running three different trades and calling it one.</p><p><strong>How to spot it:</strong> your sniper fires on most migrations of the day, and the results look random. Some trades are huge losses, some are tiny gains, and a few buys happen at prices that make no sense.</p><p><strong>The fix:</strong> decide that a migration sniper buys organic migrations only, and use the filters in the next three mistakes to enforce it. Those filters can all be checked before the buy goes out.</p><figure><img alt="Chart titled Three kinds of Pump.fun migration. Of 247 SOL-paired Pump.fun migrations to PumpSwap traced from November 3 to 5, 2026, 121 (49%) were instant: created and graduated within 5 seconds, usually by the creator, who holds most of the supply, skipped by a 10% dev cap and a 35% top 10 holder cap. 75 (30%) were Mayhem pools that opened with a median 0.21 SOL and traded at 0.13x after a minute, skipped by a 20 SOL minimum liquidity. 51 (21%) were organic, with a curve filled by real trading" src="https://cdn-images-1.medium.com/max/1024/1*mKVfRmpKrl_shyyl3NLAXA.png" /></figure><h3>Mistake 2: No minimum liquidity, so the bot buys Mayhem pools</h3><p>A normal Pump.fun migration opens a pool with dozens of SOL in it. A Mayhem migration does not. In our sample, the median Mayhem pool opened with 0.21 SOL of real liquidity. 63 of the 75 opened with less than 1 SOL, and only one had more than 10.</p><p>A pool with a fifth of a SOL in it cannot take even a small sell without the price collapsing. And the price did collapse. Measured against the first trades in each pool, the median Mayhem pool was at 0.13x after one minute and 0.12x after five. Sniper-sized buyers mostly knew this already: only one Mayhem pool in our sample saw any buy of 0.01 SOL or more in its opening block.</p><p>Mayhem coins also carry a flag on their curve and their pool. Pump’s own documentation treats them differently, for example by leaving them out of synthetic migration. A careful bot could read the flag. You do not need it to, though, because the liquidity alone gives them away.</p><p><strong>How to spot it:</strong> some buys go through at a normal size, and the sell returns almost nothing.</p><p><strong>The fix:</strong> set a minimum liquidity on the pool. At 20 SOL, it would have skipped 74 of the 75 Mayhem pools in our sample and none of the organic ones. Make sure the bot reads the SOL in the pool when the pool opens, not some number from an index that updates a minute later.</p><h3>Mistake 3: No dev holding cap, so the bot buys coins the creator bought out</h3><p>Instant migrations pass a liquidity filter. The creator paid for the whole curve, so the pool opens with normal reserves. What gives them away is who holds the tokens.</p><p>When a wallet buys an entire curve in the creation transaction, that wallet holds most of the supply outside the pool. In 96 of the 121 instant migrations in our sample, the creator was also the wallet that opened the pool. They are the market, and they can sell into your buy whenever they like.</p><p>These pools are busy. The median instant pool saw 1,408 transactions in its first five minutes, against 440 for an organic one, and its price five minutes in was 1.68 times the first trades. That looks like the best trade on PumpSwap. It was not, for small buyers. We traced 254 small first-block positions in instant pools to the moment they sold. Before costs the median was +5.1%. After network fees, priority fees and tips it was +1.9%, and together they lost 1.0 SOL on 62.9 SOL bought. The price chart went up. The wallets that bought small did not get much of that.</p><p><strong>How to spot it:</strong> the coin’s creator, or a wallet funded by the creator, holds a large share of supply right after migration.</p><p><strong>The fix:</strong> a dev holding cap. 10% is a sensible start for migration snipes. Pair it with a “skip bundled launches” check, which catches creators who split the curve buy across several wallets so a single-wallet cap does not see it. Filters like these are the subject of <a href="https://medium.com/@dexsnipecom/pump-fun-sniper-bot-filters-we-checked-807-new-coins-to-see-which-rules-work-5c7a932f9e9e">our test of Pump.fun sniper bot filters on 807 new coins</a>, if you want to see how each one behaves on fresh launches too.</p><h3>Mistake 4: No top 10 holder cap, so the bot buys after a whale opening</h3><p>This is the strangest thing in PumpSwap right now, and the most expensive for a bot that does not check.</p><p>In 35 of the 121 instant migrations we traced, one wallet bought more than half of the new pool’s tokens in the same block the pool opened. The median buy was about 2,400 SOL. The largest was 14,940 SOL. Here is one, step by step:</p><ol><li>15:30:11 UTC, November 4, 2026. A wallet creates a coin and buys its entire curve for 86 SOL in the same transaction.</li><li>The same second, the PumpSwap pool opens with the usual reserves.</li><li>In the same block, a second wallet spends 4,976 SOL and receives about 98% of the pool’s tokens.</li><li>Three minutes later, that wallet starts sending identical slices of 305,000 tokens to hundreds of other wallets, several per second.</li></ol><p>On a constant product pool, buying 98% of the tokens multiplies the price by roughly 2,500. A coin valued at about 400 SOL when the pool opened was, on paper, worth more than a million SOL a moment later, and minutes later it had hundreds of holders. All the SOL behind that price came from the wallet that holds the tokens.</p><p>We cannot see why anyone does this. It does not matter for a sniper. The price after one of these buys is not set by a market, and a bot that buys after it pays thousands of times the opening price for a token that sits almost entirely with one group. None of the 35 happened on an organic migration.</p><p><strong>How to spot it:</strong> your buy fills at a price thousands of times above the pool’s opening price, on a coin with suddenly hundreds of holders.</p><p><strong>The fix:</strong> a top 10 holder cap, checked at migration. Right after a whale opening, one wallet holds almost every token outside the pool, which no sensible cap allows. 35% is a reasonable start. A max price change since pool open, if your bot has one, catches the same thing from the other side.</p><h3>Mistake 5: Thinking “first in the pool” means “first buyer”</h3><p>Snipers talk about the first block as if it were the whole game. On PumpSwap in late 2026 it is often not the first chance to buy at all.</p><p>In October 2026, Pump documented a change to how a curve ends. Before, a buy could take at most what was left on the curve. Now, with the v3 buy instructions, the buy that completes a curve has no maximum size. It takes what is left at the curve price, completes the curve, and then keeps buying from the tokens that would have gone into the pool, at the price the pool would charge. Pump calls this a synthetic migration. When the pool opens, it opens at the price that buy left behind.</p><p>The curve records the size of that extra part, so it can be read for every coin. In 26 of the 51 organic migrations in our sample, someone had bought into the pool this way before it opened. The median was about 0.47 SOL and the largest 3.3 SOL. Mayhem coins never had one, which matches Pump’s documented rule.</p><p>The amounts are small, so the price effect is small too. The lesson is about expectations. Half the time, the true first buyer of an organic pool is whoever completed the curve, and they paid the opening price before your bot could even see the pool. Paying a large tip to be first in the opening block buys you second place.</p><p><strong>The fix:</strong> stop measuring your sniper by whether it lands in slot zero. Measure it by what it nets after costs. That leads straight to mistake 6.</p><h3>Mistake 6: Paying a big tip for a race that is not crowded</h3><p>On organic migrations, the opening block was quiet. We counted buys of 0.01 SOL or more. The median organic pool had two of them in its opening block. In the first three slots, which is well under a second, the median pool saw five buying wallets and about 2.1 SOL in total. 20 of the 51 organic pools had no buy of that size in their opening block at all.</p><p>A sniper with a normal priority fee and a modest Jito tip lands in that first second without trying hard. What a large tip buys is a few hundred milliseconds that the price does not reward. Organic pools moved a median 1% in the first ten seconds.</p><p>The cost side is not small. We followed 175 first-block buyers on organic pools, each buying 0.01 to 5 SOL, until they sold. Their network fees, priority fees and tips took a median 3.1% of the buy size. A tenth of them paid more than 15%, and those heavy tippers are where most of the gap between “before costs” and “after costs” went.</p><p>Here is how those 175 buys ended, before and after costs:</p><ul><li><strong>Share in profit:</strong> 66% before costs, 57% after costs.</li><li><strong>Median result:</strong> +13% before costs, +7.4% after costs.</li></ul><p><strong>How to spot it:</strong> almost every trade is small and slightly negative, even on coins that went up.</p><p><strong>The fix:</strong> keep the priority fee around 0.001 SOL and the Jito tip around 0.002 SOL for a migration sniper, and raise them only if you see your buys landing late on organic pools. We went through how slippage, priority fee and tip interact in <a href="https://medium.com/@dexsnipecom/best-sniper-bot-settings-for-pumpfun-slippage-priority-fee-and-tip-from-807-launches-76409e7a5040">our guide to sniper bot settings for Pump.fun</a>. The same logic holds on PumpSwap, with less pressure at the front.</p><h3>Mistake 7: Counting early buyers as demand</h3><p>Many traders, and some bots, read “ten buyers in the first second” as a sign that a pool is hot. On PumpSwap, look closer before you trust that number.</p><p>In every one of the 247 pools we traced, organic, instant and Mayhem alike, at least one of the same four wallets bought within ten seconds of opening. The busiest of them was in 245 pools. All four together were in 222. Each buy was worth a few hundred-thousandths of a SOL. They do not move the price at all. They do add to the count of early buyers and early trades.</p><p>The same goes for pools where one group sends many equal buys. Count of buyers and count of trades are easy to fake. SOL volume and how the holders are spread are much harder.</p><p><strong>The fix:</strong> if your bot or your own process uses “number of early buyers” as a signal, count only buys above a floor, for example 0.01 SOL, and look at SOL in rather than trade count. Better still, rely on the holder caps and liquidity filter above, which look at where the tokens and the SOL actually are.</p><figure><img alt="Chart titled 10 PumpSwap sniper mistakes, and the fix for each. 1 firing on every migration, fix: organic migrations only. 2 no min liquidity, fix: min liquidity 20 SOL. 3 no dev holding cap, fix: dev cap 10% plus skip bundles. 4 no top 10 holder cap, fix: top 10 cap 35%. 5 thinking slot zero means first, fix: judge by net result. 6 big tip in a quiet race, fix: priority fee 0.001 SOL and tip 0.002 SOL. 7 counting buyers as demand, fix: count SOL, not wallets. 8 a +100% take-profit, fix: +20% fo" src="https://cdn-images-1.medium.com/max/1024/1*9rlJa7r5G0W-f4kNnWbMaA.png" /></figure><h3>Mistake 8: A take-profit built for a moonshot</h3><p>The take-profit people set on a migration sniper is often +100% or more, because that is what they remember from the coins that ran. On organic PumpSwap pools, the price in the first five minutes mostly goes sideways.</p><p>Median price compared with the first three slots of each pool:</p><ul><li><strong>Organic pools:</strong> 1.01x after 10 s, 1.02x after 1 min, 1.05x after 5 min.</li><li><strong>Instant pools:</strong> 1.70x after 10 s, 1.66x after 1 min, 1.68x after 5 min.</li><li><strong>Mayhem pools:</strong> 0.33x after 10 s, 0.13x after 1 min, 0.12x after 5 min.</li></ul><p>Behind that flat median is a handover. In the first five minutes, the median organic pool took in 96 SOL of buys and paid out 99 SOL of sells. Buys come from snipers, from people who missed the curve, and from BOOST, a mechanism Pump.fun added in 2026 that spends part of each migration’s liquidity on buys in the first five minutes. Sells come from people who bought on the curve and finally have somewhere deep to sell. Both sides are large and roughly equal, so the price stays close to where it opened.</p><p>A first target of +100% on that kind of pool is a lottery ticket. A first target of +20% hits often enough to matter.</p><p><strong>How to spot it:</strong> your take-profit almost never fires, and most positions end on the stop-loss or the timer.</p><p><strong>The fix:</strong> split the exit. Sell 50% at +20% and the other 50% at +50%. The first step pays for the trade on the pools that move a little. The second rides the ones that run. Put a 20% stop-loss under both; a quarter of first-block buyers we followed lost 17% or more.</p><h3>Mistake 9: No timer, so a five-minute trade turns into a bag</h3><p>First-block buying on PumpSwap is a short trade. The median first-block buyer we followed sold 130 seconds after buying. A quarter sold within a minute, and three quarters within five minutes, which is also the length of the BOOST window.</p><p>After that window, the trade you entered is over. The curve holders have mostly sold, the boost buys have stopped, and the pool has to find new buyers on its own. A position still open at that point is a different bet: a bet on the coin’s next hour, not on the migration. Some traders want that bet. Most migration snipers forget they made it.</p><p><strong>The fix:</strong> a timer exit at about five minutes. If none of the take-profit steps or the stop-loss fired by then, the bot sells. Keep dev-sell on as well. On an organic coin, the creator selling into the new pool is the clearest sign the handover is going one way. And give the sell side a looser slippage cap than the buy side. When a pool turns, every bot sells at once, and a tight cap can make your exit fail exactly when you need it.</p><p><strong>Make sure the exits run without you.</strong> PumpSwap does not hold limit orders. Every exit is your bot watching the price and sending a sell. If your bot’s exits only run while a browser tab or a laptop is open, a five-minute timer is not really a timer.</p><figure><img alt="Timeline titled The first five minutes of an organic pool, medians over 51 organic PumpSwap pools. Before the pool opens, in 26 of 51 organic migrations a synthetic buy on the curve got in first. When the pool opens there are 2 sniper-sized buys in the opening block. At 10 seconds the price is 1.01x the first slots. At 1 minute it is 1.02x and a quarter of first buyers have sold. At 130 seconds the median first-block buyer sells. At 5 minutes the price is 1.05x, the BOOST window ends and a timer" src="https://cdn-images-1.medium.com/max/1024/1*CdHuWwyQ7QhVWt2lSnzlmA.png" /></figure><h3>Mistake 10: Buying the wrong pool, or paying fees you did not plan for</h3><p>Anyone can create a PumpSwap pool for any token. The canonical pool is the one the Pump program makes at migration, but other pools for the same coin can exist, some with almost nothing in them. A bot that searches for “a PumpSwap pool with this mint” instead of deriving the canonical pool can buy the wrong one.</p><p>The canonical pool is easy to check. Its creator field is a Pump program address derived from the coin’s mint, and Pump’s documentation gives the exact derivation. A good migration sniper only buys the pool the migration created.</p><p>Fees are the other half of this mistake. A PumpSwap trade pays an LP fee, a protocol fee and, on canonical pools, a coin creator fee. Since September 2025 the fee on canonical pools comes from tiers based on the coin’s market cap, so a buy on a small coin and a buy on a larger one can pay different rates. Pump told integrators that bots which had not implemented the new fee logic should raise their slippage tolerance as a temporary fix. If your bot still computes fees the old way, some of its “slippage” fails are really fee math.</p><p>Then there is the bot’s own fee. Many Telegram and web bots take a share of every trade, often around 1%. On a trade whose median result is +7.4% after network costs, a 1% cut on the buy and another on the sell takes a big part of the edge. Check what your bot charges per trade before you check anything else.</p><p><strong>The fix:</strong> use a bot that buys only the canonical pool, computes PumpSwap fees from the current tiers, and either charges nothing per trade or charges an amount you have priced into your take-profit.</p><h3>The PumpSwap sniper settings that avoid all ten</h3><p>Here is a starter setup built from the mistakes above. Treat it as a starting point and change one thing at a time once you have your own results.</p><ul><li><strong>Trigger:</strong> Migration, Pump.fun only. Fixes mistake 10 (buys only the canonical pool).</li><li><strong>Buy size:</strong> 0.2 to 0.5 SOL. Fixes mistake 6 (costs stay a small share).</li><li><strong>Buy slippage:</strong> 10 to 15%. Fixes mistake 10 (room for fee tiers).</li><li><strong>Priority fee:</strong> 0.001 SOL. Fixes mistake 6.</li><li><strong>Jito tip:</strong> 0.002 SOL. Fixes mistakes 5 and 6.</li><li><strong>Min liquidity:</strong> 20 SOL. Fixes mistake 2 (skips Mayhem pools).</li><li><strong>Dev holding cap:</strong> 10%. Fixes mistake 3 (skips creator buyouts).</li><li><strong>Top 10 holder cap:</strong> 35%. Fixes mistake 4 (skips whale openings).</li><li><strong>Skip bundled launches:</strong> On. Fixes mistake 3.</li><li><strong>Take-profit:</strong> +20% sell 50%, +50% sell 50%. Fixes mistake 8.</li><li><strong>Stop-loss:</strong> 20%. Fixes mistake 8.</li><li><strong>Timer:</strong> 5 minutes. Fixes mistake 9.</li><li><strong>Dev-sell exit:</strong> On. Fixes mistake 9.</li><li><strong>Sell slippage:</strong> Looser than buy, 20 to 25%. Fixes mistake 9.</li></ul><p>Two settings are missing on purpose. There is no “min early buyers” filter, because of mistake 7. And there is no max market cap, because organic migrations all open at roughly the same size; the filters above already separate them from the rest.</p><figure><img alt="Settings card titled A starter PumpSwap migration sniper. Buy: Migration trigger, Pump.fun venue, size 0.2 to 0.5 SOL, slippage 10 to 15%, priority fee 0.001 SOL, Jito tip 0.002 SOL. Filters: min liquidity 20 SOL, dev holding cap 10%, top 10 holders 35%, skip bundles on, mint and freeze check on. Exits: take-profit 1 at +20% selling 50%, take-profit 2 at +50% selling 50%, stop-loss 20%, timer 5 minutes, dev-sell on, sell slippage 20 to 25%. Note: on organic pools first-block buyers made a median" src="https://cdn-images-1.medium.com/max/1024/1*A_J1qbtdkzxX8VjLVNIM_w.png" /></figure><h3>How to set up a PumpSwap migration sniper, step by step</h3><ol><li><strong>Make a separate sniper wallet.</strong> Create a new Solana wallet and send it only the SOL you plan to trade. Your main wallet never signs a sniper trade. Never give any bot a seed phrase.</li><li><strong>Create a new sniper and choose the Migration trigger.</strong> Tick Pump.fun as the venue, so the bot buys the PumpSwap pool the migration creates and nothing else.</li><li><strong>Set the buy.</strong> 0.2 to 0.5 SOL, buy slippage 10 to 15%, priority fee 0.001 SOL, Jito tip 0.002 SOL.</li><li><strong>Set the filters.</strong> Min liquidity 20 SOL, dev holding cap 10%, top 10 holder cap 35%, skip bundled launches on. If your bot has mint and freeze authority checks, leave them on too.</li><li><strong>Build the exits.</strong> Two take-profit steps (+20% sell half, +50% sell the rest), a 20% stop-loss, a 5 minute timer, dev-sell on, sell slippage 20 to 25%.</li><li><strong>Arm it with a small size first.</strong> Let it take ten trades at the smallest size you are comfortable with.</li><li><strong>Read the ten trades.</strong> For each one, check which kind of migration it was, what you paid in fees and tips, and which exit closed it. If a trade was an instant or Mayhem pool, a filter is too loose.</li><li><strong>Change one thing at a time.</strong> If buys land late on organic pools, raise the tip a little. If take-profits never fire, lower the first step. If sells fail, loosen sell slippage.</li><li><strong>Review weekly.</strong> The mix of migration kinds moves from week to week. The signatures that tell them apart change much more slowly.</li></ol><h3>What to look for in a PumpSwap sniper bot</h3><p>If you are choosing a bot rather than tuning one, these questions sort them quickly:</p><ul><li><strong>Does it buy only the canonical pool?</strong> Ask how it finds the pool. “It derives the Pump pool from the mint” is the right answer.</li><li><strong>Can it check holders at migration?</strong> A dev holding cap and a top 10 cap, both read when the pool opens, are what skip instant and whale-bought pools.</li><li><strong>Does its liquidity filter read the pool itself?</strong> A filter that waits for a third-party index is too slow for a pool that trades in its first second.</li><li><strong>Do exits run on a server?</strong> A timer is only useful if it fires when your laptop is closed.</li><li><strong>Does it take a share of every trade?</strong> On a trade with a single-digit median edge, a per-trade cut can be most of it.</li><li><strong>Is it current with PumpSwap fee tiers and the v3 curve instructions?</strong> Pump changes its programs often. A bot that still uses old fee math fails more sells than it should.</li></ul><p>Migration sniping is one way to trade Pump.fun coins. If you want to compare it with buying at launch, our trace of <a href="https://medium.com/@dexsnipecom/solana-sniper-bot-who-really-buys-first-we-traced-411-early-pumpfun-buys-6cc602a342c2">who really buys first on Pump.fun</a> shows what the front of the curve looks like. The same lifecycle on Raydium’s side is in <a href="https://medium.com/@dexsnipecom/letsbonk-sniper-bot-how-to-snipe-bonk-fun-coins-at-every-stage-from-launch-to-raydium-0a785acffa8c">our LetsBonk sniper bot guide</a>, where coins graduate to a Raydium pool instead of PumpSwap.</p><h3>When a migration sniper is the wrong tool</h3><p>A migration sniper is good at one thing: a short trade on the handover from curve holders to new buyers. It is the wrong tool if:</p><ul><li><strong>You want to hold for days.</strong> The filters here are tuned for the first five minutes. A coin’s next week depends on things no migration filter can see.</li><li><strong>You want big multiples.</strong> Organic pools mostly traded sideways in their first minutes. The large multiples on Pump.fun come earlier, on the curve, with a much higher chance of going to zero.</li><li><strong>You cannot accept many small losses.</strong> Even on the best kind of migration, a third of first-block buyers lost money before costs, and more than four in ten after costs.</li><li><strong>You only trade a handful of coins.</strong> The edge, where there is one, shows up over many trades. Ten trades is a test, not a result.</li></ul><h3>Questions people ask about PumpSwap sniper bots</h3><h4>What is a PumpSwap sniper bot?</h4><p>A bot that buys a Pump.fun coin on PumpSwap in the first blocks after its bonding curve completes and its pool opens, then sells it on rules you set, such as take-profit steps, a stop-loss and a timer.</p><h4>How do you snipe a Pump.fun migration?</h4><p>Use a migration trigger that watches for completed Pump.fun curves, derive the canonical PumpSwap pool, run filters on that pool (minimum liquidity, dev holding cap, top 10 holder cap), and send the buy with a modest priority fee and tip. Set your exits before you arm it.</p><h4>When does a Pump.fun coin migrate to PumpSwap?</h4><p>When its bonding curve is complete, meaning enough SOL has gone into the curve. After that, no one can trade on the curve, and any wallet can call the migrate instruction to open the PumpSwap pool. Sometimes this happens seconds after creation, when the creator buys the whole curve at once.</p><h4>What is a synthetic migration on Pump.fun?</h4><p>Since October 2026, the buy that completes a Pump.fun curve can be larger than what is left on it. The extra part is priced as if the PumpSwap pool already existed, and the pool opens at the price that buy left behind. Mayhem coins are excluded.</p><h4>What slippage should I use for PumpSwap sniping?</h4><p>For organic migrations, 10 to 15% on the buy is a reasonable start, and a looser 20 to 25% on the sell so exits do not fail when many bots sell at once. If your bot has not been updated for PumpSwap’s market cap fee tiers, it may need more.</p><h4>What is a Mayhem pool on PumpSwap?</h4><p>The pool of a coin launched with Pump.fun’s Mayhem Mode. In our sample, Mayhem pools opened with a median of about 0.2 SOL of real liquidity and fell to around an eighth of their first trades within a minute. A minimum liquidity filter of 20 SOL skips them.</p><h4>Does a PumpSwap sniper bot need Jito?</h4><p>It helps your buy land reliably, but the tip can be small. Organic pools had a median of two sniper-sized buys in their opening block, so you do not need to outbid a crowd. Large tips mostly reduced what first-block buyers kept.</p><h4>How long should I hold a migration snipe?</h4><p>Most first-block buyers we followed sold within five minutes, and the median sold after 130 seconds. A timer of about five minutes matches the trade, unless you mean to make a longer bet on purpose.</p><h4>Are PumpSwap sniper bots safe?</h4><p>The bot itself is only as safe as its wallet setup. Use a separate sniper wallet that holds only trading money, never share a seed phrase, and prefer bots that sign in with your main wallet read-only. The trade itself is high risk: most meme coins lose value.</p><h3>How we know these numbers</h3><p>The migration figures come from our own trace of Pump.fun migrations. We listed the transaction history of an account that takes part in every Pump.fun migration for November 3 to 5, 2026, picked migrations at random and kept 291 distinct coins. 44 were paired with something other than SOL and left out, leaving 247. For each coin we derived its canonical PumpSwap pool and its bonding curve, read the curve’s Mayhem flag, quote mint and synthetic migration fields, and took the curve’s oldest transaction as the creation time. An instant migration is one whose pool opened within five seconds of creation. A whale opening is a single buy in the opening block of more than half the pool’s starting tokens.</p><p>For every pool we read all successful transactions in the first five minutes and measured each trade from the change in the pool’s own token and SOL vaults. For the results, we took every buy of 0.01 to 5 SOL in the first three slots of a pool, kept wallets that paid in plain SOL and never moved the tokens elsewhere, and followed each one’s token account until it sold. “Before costs” is SOL paid into and taken out of the pool. “After costs” is the change in the wallet’s SOL balance, which also includes network fees, priority fees, tips and token account rent.</p><p>Three days of migrations is a snapshot. The mix of instant, Mayhem and organic migrations will be different in other weeks. The ways to tell them apart, and the filters that do it, change much more slowly. The starter settings are built from these numbers; they are starting points, not promises. The program rules come from Pump’s public documentation, checked on November 13, 2026.</p><h3>Sources and further reading</h3><ul><li>Pump public docs, synthetic migration (the completing buy, its rules, and the Mayhem exception): <a href="https://github.com/pump-fun/pump-public-docs/blob/main/docs/SYNTHETIC_MIGRATION.md">github.com/pump-fun/pump-public-docs</a></li><li>Pump public docs, fee program (market cap fee tiers for curves and canonical PumpSwap pools, from September 2025): <a href="https://github.com/pump-fun/pump-public-docs/blob/main/docs/FEE_PROGRAM_README.md">github.com/pump-fun/pump-public-docs</a></li><li>Pump public docs, PumpSwap coin creator fee (canonical pool definition and the creator fee on each swap): <a href="https://github.com/pump-fun/pump-public-docs/blob/main/docs/PUMP_SWAP_CREATOR_FEE_README.md">github.com/pump-fun/pump-public-docs</a></li><li>Pump public docs, PumpSwap program README (pool layout and flags): <a href="https://github.com/pump-fun/pump-public-docs/blob/main/docs/PUMP_SWAP_README.md">github.com/pump-fun/pump-public-docs</a></li><li>Pump program IDL (migrate instruction and the BondingCurve layout): <a href="https://github.com/pump-fun/pump-public-docs/blob/main/idl/pump.json">github.com/pump-fun/pump-public-docs</a></li><li>PumpSwap program on Solscan: <a href="https://solscan.io/account/pAMMBay6oceH9fJKBRHGP5D4bD4sWpmSwMn52FMfXEA">solscan.io</a></li><li>The 4,976 SOL opening buy described above, on Solscan: <a href="https://solscan.io/tx/3ZfNX5a4wY277H6eXwqpQ9nsPusRd4GJPat4xj2TkMCN2cPgt8jBStSsBhV8jWo2o8MUupGY9Gr9Dzbj8Skys4jf">solscan.io</a></li><li>Decrypt, Pump.fun debuts PumpSwap exchange (March 2025): <a href="https://decrypt.co/310933/solana-pump-fun-pumpswap-exchange">decrypt.co</a></li><li>SolanaFloor, Pump.fun unleashes Mayhem Mode (November 2025): <a href="https://solanafloor.com/news/one-more-effort-to-revive-struggling-solana-trenches-pump-fun-unleashes-mayhem-mode">solanafloor.com</a></li><li>The Block, Pump.fun graduation rate jumps after BOOST changes launch incentives (July 2026): <a href="https://www.theblock.co/news/defi/2026-07-29-pump-fun-token-graduation-rate-jumps-boost-changes-launch-incentives-409815">theblock.co</a></li><li>Jito docs, tips and bundles on Solana: <a href="https://docs.jito.wtf/lowlatencytxnsend/">docs.jito.wtf</a></li></ul><p>Nothing here is financial advice. Meme coins are very risky, and most new tokens go to zero. Only trade with money you can afford to lose.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=871f672d0819" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[LetsBonk Sniper Bot: How to Snipe Bonk.fun Coins at Every Stage, From Launch to Raydium]]></title>
            <link>https://medium.com/@dexsnipecom/letsbonk-sniper-bot-how-to-snipe-bonk-fun-coins-at-every-stage-from-launch-to-raydium-0a785acffa8c?source=rss-2654b58e3d1e------2</link>
            <guid isPermaLink="false">https://medium.com/p/0a785acffa8c</guid>
            <category><![CDATA[crypto-trading]]></category>
            <category><![CDATA[trading-bot]]></category>
            <category><![CDATA[memecoins]]></category>
            <category><![CDATA[cryptocurrency]]></category>
            <category><![CDATA[solana-blockchain]]></category>
            <dc:creator><![CDATA[DexSnipe]]></dc:creator>
            <pubDate>Fri, 13 Nov 2026 05:06:12 GMT</pubDate>
            <atom:updated>2026-11-13T05:06:12.575Z</atom:updated>
            <content:encoded><![CDATA[<h3>A LetsBonk coin gives a sniper five different moments to act. Here is what happens at each one, which trigger fits it, and the settings and exits to use.</h3><p>Most guides treat sniping as one thing: be fast, buy the new coin, hope. On LetsBonk it is more useful to think of a coin’s life as a short trip with five stops. There is the time before it exists, when only its creator knows it is coming. There is block zero, when the curve opens. There is the curve itself, where the price climbs with every buy. There is graduation, when the curve hits its target and the coin moves to a Raydium pool. And there is life after that, in a normal pool where anyone can trade. A LetsBonk sniper bot can act at each of these stops, but the right trigger, the right size and the right exits are different every time. This guide walks the trip in order, then shows you how to set up a sniper from zero, with three starting presets you can copy.</p><figure><a href="https://www.dexsnipe.com/"><img alt="DexSnipe banner: One LetsBonk coin, five moments to act. A dashboard illustration shows four snipers: creator watch on LetsBonk plus Pump.fun and LaunchLab, fresh launch with SOL pairs and a 10% dev cap, migration with a 50 SOL minimum, and copy trade, with exits on every sniper. The button links to https://www.dexsnipe.com/" src="https://cdn-images-1.medium.com/max/1024/1*93vr8ceiaFiX6STKfzHmiw.png" /></a><figcaption>DexSnipe is our own Solana sniper bot service.</figcaption></figure><h3>The short version</h3><ul><li><strong>LetsBonk runs on Raydium LaunchLab.</strong> Every LetsBonk coin starts on a LaunchLab bonding curve. On a standard SOL curve, 793.1 million of the 1 billion tokens are for sale, and the curve graduates when it collects its target, 85 SOL by default. Then the coin moves to a Raydium CPMM pool.</li><li><strong>Pick the moment before you pick the bot.</strong> Creator watch acts before launch. Fresh launch acts at block zero. Copy trade acts on the curve. Migration acts at graduation. Exits do the work after that.</li><li><strong>Speed matters most at block zero.</strong> On a standard curve, a 0.5 SOL buy gets 1.77 times worse price if 10 SOL bought ahead of you, and 2.76 times worse if 20 SOL did.</li><li><strong>Most coins never leave the curve.</strong> When we counted, about 1.1% of LetsBonk launches had graduated. A sniper that buys blind needs small size and fast exits.</li><li><strong>Some graduations are not real.</strong> We found 777 coins whose creator bought the whole curve in the same transaction that created it. A migration sniper needs a liquidity floor.</li><li><strong>Exits are the bot watching and selling.</strong> A curve does not hold limit orders. Every take profit and stop loss is your bot sending a sell at the right time.</li></ul><h3>What is a LetsBonk sniper bot?</h3><p>LetsBonk (you will also see it called Bonk.fun) is a meme coin launchpad on Solana from the BONK community. It went live in April 2025, and by July 2025 it was creating more new tokens per day than Pump.fun for the first time. Under the hood it is not its own program. It is a platform on Raydium LaunchLab, the launch program that Raydium opened to third parties. Each LetsBonk coin is a LaunchLab curve that records LetsBonk as the platform that made it.</p><p>A LetsBonk sniper bot is a program that watches that curve program and buys LetsBonk coins at a moment you choose, then sells them on rules you set. “A moment you choose” is the part people skip. A sniper can buy:</p><ul><li>the instant a wallet you follow creates a new coin,</li><li>in the first block of any new LetsBonk coin that passes your filters,</li><li>when a wallet you copy buys a coin on the curve,</li><li>or the moment a coin graduates and its Raydium pool opens.</li></ul><p>Those are four different trades with four different risks. A bot that does one well can do another badly. So before we talk about settings, let us walk through a coin’s life and see what each moment looks like from the sniper’s side.</p><h4>How a LetsBonk coin works, in one paragraph</h4><p>When a creator launches, LaunchLab mints the token, puts the part that is for sale into a vault, and opens a curve priced in a quote coin, usually SOL. Buyers send SOL to the curve and get tokens. Every buy raises the price, every sell lowers it. The SOL that buyers pay piles up in a second vault. When that pile reaches the graduation target, the program takes a snapshot, creates a Raydium CPMM pool with the SOL and the tokens that are left, revokes the token’s mint authority so no more can ever be made, and closes the curve. From then on, the coin trades in that pool like any other Raydium token.</p><h3>The five moments of a LetsBonk coin</h3><figure><img alt="Timeline of the five moments of a LetsBonk coin and the sniper trigger for each: 1 before launch, creator watch; 2 block zero, fresh launch; 3 on the curve, copy trade; 4 graduation at 85 SOL into a Raydium CPMM pool, migration; 5 after graduation, your exits. About 1.1% of 1,150,183 LetsBonk launches had graduated when DexSnipe counted in November 2026." src="https://cdn-images-1.medium.com/max/1024/1*N_f5AQQLdiKdTifL37JdcQ.png" /></figure><p>Here is the trip in order, with the trigger that fits each stop. We will go through them one by one.</p><ol><li><strong>Before launch.</strong> The coin does not exist yet. Only the creator knows it is coming. Trigger: creator watch.</li><li><strong>Block zero.</strong> The curve opens and the first buys land. Trigger: fresh launch.</li><li><strong>On the curve.</strong> The price climbs as buyers come in. Trigger: copy trade, or a fresh launch sniper that is late.</li><li><strong>Graduation.</strong> The curve hits its target and a Raydium pool opens. Trigger: migration.</li><li><strong>After graduation.</strong> A normal pool, open to every wallet and every router. No new trigger. This is where your exits do their job.</li></ol><h3>Moment 1: before launch, and why creator watch fits LetsBonk</h3><p>The cleanest snipe is the one that starts before the coin exists. A creator watch sniper follows one wallet. When that wallet creates a new coin, the sniper buys it in the same moment, with no need to guess which of thousands of new coins is worth a look.</p><p>This works on LetsBonk because so many coins come from repeat creators. When we matched 1,150,183 launches on LetsBonk’s three main platform accounts to their creators, we found 218,856 wallets. Most launched once. But 144 wallets had launched 1,000 coins or more each, and together they made about 30% of everything on LetsBonk. Those mass launchers are not the ones to follow. Their coins graduate less often than average, because they are playing a numbers game.</p><p>The ones to follow are smaller creators with a record. In a time-ordered sample of 220 creators, the launches that came after a creator’s first graduation graduated 4.6% of the time. The LetsBonk average was about 1.1%. That is roughly four times better odds, and it is the whole reason creator watch exists.</p><h4>How to pick a creator to watch</h4><ol><li><strong>Start from coins you know.</strong> Take a LetsBonk coin that graduated and did well. Open it on Solscan and find the wallet that created it.</li><li><strong>Read the wallet’s history.</strong> How many coins has it launched? How many graduated? A wallet with 30 launches and 3 graduations is a better bet than one with 3,000 launches and 10.</li><li><strong>Check the pace.</strong> In our sample, the median gap between one launch and the next was about 75 minutes, a quarter came within 7 minutes, and three quarters within a day. A creator who launched last week and not since may have moved on.</li><li><strong>Check where else it launches.</strong> A creator wallet is just a Solana address. Nothing stops it from launching its next coin on Pump.fun or another LaunchLab platform. Your creator watch should cover those venues too, or you will miss the launch.</li><li><strong>Keep the list short and prune it.</strong> Two to five creators is plenty. If a creator launches ten more times with no graduation, remove them.</li></ol><h4>What can go wrong with creator watch</h4><p>A creator who sees snipers follow every launch can use them. The simple version: take a big first buy, then sell it into the snipers’ demand in the first minutes. That is why a creator watch sniper still needs a dev holding cap (skip the coin if the creator’s own buy is too large a share of supply) and a dev sell exit (sell if the creator starts selling). You bought because of who launched the coin. If that wallet is selling, the reason is gone.</p><h3>Moment 2: block zero, where your price is decided</h3><p>The second moment is the launch itself. A fresh launch sniper does not know the creator. It sees every new LetsBonk coin, runs your filters, and buys the ones that pass, as close to the first block as it can.</p><p>Why the rush? Because on a bonding curve, your price is set by how much SOL went in before you. Here is the same 0.5 SOL buy on a standard LetsBonk SOL curve, at different points in the queue (before fees):</p><figure><img alt="Bar chart of the same 0.5 SOL buy on a standard LetsBonk SOL curve, before fees: first buyer gets 17.59 million tokens at 1.00x price, 2 SOL ahead 15.48 million at 1.14x, 10 SOL ahead 9.94 million at 1.77x, 20 SOL ahead 6.37 million at 2.76x, 40 SOL ahead 3.26 million at 5.39x. The curve starts with 30 SOL of virtual SOL and graduates at 85 SOL." src="https://cdn-images-1.medium.com/max/1024/1*TnySU3Zn3ofCEX9YxGhA9w.png" /></figure><ul><li><strong>You are first:</strong> about 17.59 million tokens. Price 1.00x.</li><li><strong>2 SOL bought ahead of you:</strong> about 15.48 million tokens. Price 1.14x.</li><li><strong>10 SOL ahead:</strong> about 9.94 million tokens. Price 1.77x.</li><li><strong>20 SOL ahead:</strong> about 6.37 million tokens. Price 2.76x.</li><li><strong>40 SOL ahead:</strong> about 3.26 million tokens. Price 5.39x.</li></ul><p>The math behind this is simple. A standard LetsBonk curve acts like a pool that starts with 30 SOL of “virtual” SOL on one side. Every real SOL you add makes the next token more expensive. With 30 SOL of padding, the first few SOL barely move the price. By the time 20 or 40 SOL has gone in, the same 0.5 SOL buys a fraction of what it bought at the start.</p><p>So at block zero, two settings matter more than anything else:</p><ul><li><strong>Priority fee and Jito tip.</strong> These decide where your buy lands in the block. A sniper that tips too little lands behind the ones that tipped more.</li><li><strong>Slippage cap.</strong> This is the most you will accept above the price you expected. Too tight and your buy fails when others land first. Too loose and you buy at a price you would never have chosen.</li></ul><p>We went deep on how to set these in our guide to <a href="https://medium.com/@dexsnipecom/best-sniper-bot-settings-for-pumpfun-slippage-priority-fee-and-tip-from-807-launches-76409e7a5040">the best sniper bot settings for slippage, priority fee and tip</a>. That data comes from Pump.fun launches, but the curve math is the same idea, and the lessons carry over.</p><h4>Filters that matter at block zero on LetsBonk</h4><p>A fresh launch sniper buys blind into a market where most coins never graduate, so filters do the heavy lifting. These are the ones that matter most on LetsBonk:</p><ol><li><strong>Dev holding cap.</strong> Skip coins where the creator’s own first buy is already a big share of supply. On a constant product curve that share is cheap, and it is usually the first thing sold.</li><li><strong>Skip bundled launches.</strong> Some creators buy their own coin from many wallets in the same block. That hides how much they really hold. A bundle check catches it.</li><li><strong>Top 10 holder cap.</strong> Catches concentration the dev cap misses, like a creator who sent tokens to friends.</li><li><strong>Quote pair.</strong> Most LetsBonk coins pair with SOL. Some pair with USD1, BONK, TRUMP and other coins, and every price and target on those curves is in that coin. Decide which pairs you want and skip the rest.</li><li><strong>Mint and freeze authority.</strong> On a LaunchLab curve the program holds the mint authority until graduation, so this check matters most for migration snipes and for copycats that are not real LaunchLab coins.</li><li><strong>Blocklist words.</strong> Copycat names and tickers of whatever is trending today make up a big part of mass launcher output. A short word list keeps your sniper out of them.</li></ol><p>If you want to see how filters like these performed on real coins, we tested them in <a href="https://medium.com/@dexsnipecom/pump-fun-sniper-bot-filters-we-checked-807-new-coins-to-see-which-rules-work-5c7a932f9e9e">Pump.fun sniper bot filters</a>, where we followed 807 new coins for 24 hours.</p><h3>Moment 3: on the curve, and the trade fee nobody counts</h3><p>The third moment is the stretch between launch and graduation. Most LetsBonk coins live and die here. The price goes up as buyers come in and down as they sell. There is no pool yet, just the curve.</p><p>A sniper can act here in two ways. A fresh launch sniper that was slow ends up here by accident. A copy trade sniper ends up here on purpose: it follows a trader’s wallet and buys when they buy. Copy trading on the curve can work if the wallet you follow is early and good. But you are always behind them, so you always pay more than they did. Look at the queue numbers above. If the trader you copy bought with 2 SOL ahead and you land with 10 SOL ahead, you paid more than 1.5 times their price for the same coin.</p><h4>The fee on every curve trade</h4><p>Every buy and every sell on a LaunchLab curve pays a trade fee. Raydium’s docs describe up to four parts: a protocol fee for Raydium, a platform fee for the launchpad, an optional creator fee, and a referral fee. The exact rates are stored on each curve. Raydium’s own worked example uses 1% on buys and 1% on sells.</p><p>That sounds small, but a sniper trades in and out fast. At 1% each way, a round trip costs about 2% before network fees, priority fees and tips. A take profit at +20% is really closer to +18%. A sniper that flips coins for small gains many times a day pays this fee every time. Count it when you set your first take profit.</p><h4>When to sell on the curve</h4><p>Most coins that pump on a curve give it back within the hour. The simplest rule that works: make your first take profit step sell enough to get your cost back. After that, whatever is left is a free ride. If the coin keeps climbing toward graduation, your later steps catch it. If it fades, you already have your SOL.</p><h3>Moment 4: graduation, the 85 SOL line</h3><p>The fourth moment is when the curve’s SOL vault reaches the graduation target. On a standard LetsBonk SOL curve that target is 85 SOL. Raydium’s docs list four things that happen when it is reached:</p><ol><li>The program takes a snapshot of the tokens and SOL left on the curve.</li><li>It creates a Raydium CPMM pool seeded with those reserves.</li><li>It revokes the token’s mint authority, so no more can ever be minted.</li><li>It marks the curve as graduated. After that, buys and sells on the curve are rejected.</li></ol><p>For launches that migrate after Raydium’s August 17, 2026 upgrade, the docs say the platform and creator share of the new pool’s LP is locked, and the rest is burned. In plain words: the liquidity that goes into the pool cannot be pulled by the creator. That removes one of the classic ways a new pool goes to zero.</p><h4>What a migration sniper does</h4><p>A migration sniper waits for graduation and buys the moment the Raydium pool opens. The logic is that graduation is proof of demand. 85 SOL of real buying has already happened, and now the coin is in a pool that every router and aggregator can see. That brings in a new wave of buyers who never trade on curves.</p><p>The trade off: you are not early. The curve buyers are already up, and some of them will sell into the new pool. A migration snipe works best with a moderate size, a tighter slippage cap than a launch snipe (the pool is deeper, so price moves less per buy), and a trailing stop.</p><h4>The graduation trap: coins that graduate in one transaction</h4><p>Not every graduation is real demand. When we read the history of every graduated LetsBonk curve, we found 777 that had exactly two transactions: the creation, which already bought the entire curve, and the migration. No one else ever had a chance to buy. These came from just 31 wallets.</p><p>757 of the 777 were paired with the TRUMP token, not SOL, and most had graduation targets under half a TRUMP. So the “graduated” pool held almost nothing. A migration sniper with no liquidity floor would buy into a pool where one small sell moves the price off a cliff.</p><p>The fix is one setting: a minimum liquidity for migration snipes, in SOL. Set it, and these instant graduations simply do not trigger your sniper.</p><h3>Moment 5: after graduation, where exits do the work</h3><p>After graduation, the coin is in a normal Raydium CPMM pool. Trades there pay the pool’s fee, which Raydium’s docs say is typically the 0.25% tier for migrated pools. There is no new trigger at this stage. This is where the exits you set earlier earn their keep.</p><p>A curve or pool does not hold your orders. There is no “limit sell” sitting on-chain. Every exit is your bot watching the price and sending a sell when a rule is hit. That has two consequences:</p><ul><li><strong>The bot has to be running.</strong> If your sniper only runs while a browser tab is open, your stop loss stops when you close the laptop. Exits need to run on a server, all the time.</li><li><strong>Sells need slippage room too.</strong> On a falling coin, a sell with a tight slippage cap fails, and fails again, while the price keeps dropping. Give sells more room than buys.</li></ul><h4>The five exits a LetsBonk sniper should have</h4><ol><li><strong>Take profit steps.</strong> Sell part at one target, more at the next. Two to four steps is normal. The first one should cover your cost.</li><li><strong>Stop loss.</strong> Sell everything if the price drops a set percent below your entry. This caps the cost of being wrong.</li><li><strong>Trailing stop.</strong> Follows the price up and sells if it falls a set percent from the top. Best after graduation, when the pool is deep and selling is easy.</li><li><strong>Timer.</strong> Sell after a set time no matter what. Most fresh launches that will do anything do it fast. A coin that has done nothing for 20 minutes is usually done.</li><li><strong>Dev sell.</strong> Sell if the creator sells. A must on creator watch snipers.</li></ol><h3>How to set up a LetsBonk sniper bot, step by step</h3><p>Now that the moments are clear, here is the setup from zero. These steps fit any rule-based sniper. Where we name screens, they are from DexSnipe, since that is the one we know best, but the idea is the same elsewhere.</p><figure><img alt="Ten steps to set up a LetsBonk sniper bot: 1 separate wallet, 2 pick the moment, 3 tick venues, 4 trigger details, 5 buy settings such as size, slippage, priority fee and tip, 6 filters such as dev cap, bundles, top 10 and quote pair, 7 exits with a first take profit that covers your cost, 8 arm and test small, 9 change one setting at a time, 10 review weekly. Never give any bot a seed phrase." src="https://cdn-images-1.medium.com/max/1024/1*DEXz4nhwaijFuExkXD0XXg.png" /></figure><ol><li><strong>Make a separate sniper wallet.</strong> Create a new Solana wallet that you use for nothing else. Fund it with only what you plan to trade. Never use your main wallet, and never give any bot a seed phrase.</li><li><strong>Decide which moment you are trading.</strong> Creator watch, fresh launch, migration, or copy trade. Start with one. You can add more snipers later.</li><li><strong>Pick the venue.</strong> Tick LetsBonk. For creator watch, also tick the other venues the creator might use, such as Pump.fun and the rest of Raydium LaunchLab.</li><li><strong>Set the trigger details.</strong> For creator watch, paste the creator’s wallet. For copy trade, paste the trader’s wallet. For fresh launch and migration, there is nothing to paste; the filters decide.</li><li><strong>Set the buy.</strong> Buy size in SOL, slippage cap, priority fee and Jito tip. Use the presets below as a start.</li><li><strong>Set the filters.</strong> Dev holding cap, bundle check, top 10 cap, quote pair, blocklist words, and a minimum liquidity for migration snipes.</li><li><strong>Set the exits.</strong> First take profit covers your cost. Add a stop loss. Add a trailing stop for migration snipes, a timer for fresh launches, and dev sell for creator watch.</li><li><strong>Arm it and test small.</strong> Start with the smallest size you are willing to lose. Watch the first ten trades closely. Did buys land where you expected? Did sells go through?</li><li><strong>Adjust one thing at a time.</strong> If buys land late, raise the tip. If buys fail, raise the slippage cap a little. If sells fail on falling coins, raise the sell slippage. Change one setting, then watch again.</li><li><strong>Review weekly.</strong> Prune creators that stopped working, check that your filters still catch what you want, and look at whether your timer is too short or too long.</li></ol><h3>Three LetsBonk sniper presets you can start from</h3><p>These are starting points, not promises. Every market week is different, and you should tune them to your own size and risk. Each one fits a different moment.</p><h4>Preset A: creator watch (moment 1)</h4><ul><li><strong>Trigger:</strong> creator watch, one wallet per sniper</li><li><strong>Buy size:</strong> 0.25 to 0.5 SOL</li><li><strong>Slippage cap:</strong> 20 to 30%</li><li><strong>Priority fee:</strong> 0.002 SOL. <strong>Jito tip:</strong> 0.004 SOL</li><li><strong>Dev holding cap:</strong> 15%. <strong>Skip bundled launches:</strong> on</li><li><strong>Take profit:</strong> +100% sell 50%, +300% sell 30%</li><li><strong>Stop loss:</strong> 35%. <strong>Timer:</strong> off. <strong>Dev sell:</strong> on</li></ul><p>Why it is the most aggressive: it fires rarely and on purpose, so it can afford a bigger tip to land in the first block, where the queue numbers show the price edge is largest. The dev cap is a little looser because a creator you chose for their record may take a normal first buy.</p><h4>Preset B: fresh launch (moment 2)</h4><ul><li><strong>Trigger:</strong> fresh launch on LetsBonk</li><li><strong>Buy size:</strong> 0.05 to 0.15 SOL</li><li><strong>Slippage cap:</strong> 15 to 25%</li><li><strong>Priority fee:</strong> 0.001 SOL. <strong>Jito tip:</strong> 0.002 SOL</li><li><strong>Dev holding cap:</strong> 10%. <strong>Skip bundled launches:</strong> on. <strong>Blocklist:</strong> trending copycat words</li><li><strong>Take profit:</strong> +60% sell 50%, +150% sell 50%</li><li><strong>Stop loss:</strong> 30%. <strong>Timer:</strong> 20 minutes</li></ul><p>Why it is small: it buys blind into a market where about 99 coins in 100 never graduate. Small size, tight exits and a timer keep the cost of being wrong low enough that the rare winner pays for the rest.</p><h4>Preset C: migration (moment 4)</h4><ul><li><strong>Trigger:</strong> migration (graduation to Raydium)</li><li><strong>Buy size:</strong> 0.25 to 0.5 SOL</li><li><strong>Slippage cap:</strong> 10 to 15%</li><li><strong>Priority fee:</strong> 0.001 SOL. <strong>Jito tip:</strong> 0.002 SOL</li><li><strong>Minimum liquidity:</strong> 50 SOL</li><li><strong>Take profit:</strong> +40% sell 50%, +120% sell 50%</li><li><strong>Stop loss:</strong> 25%. <strong>Trailing stop:</strong> on. <strong>Timer:</strong> 60 minutes</li></ul><p>Why the slippage is tighter: the pool is deeper than a fresh curve, so each buy moves the price less. The minimum liquidity is the setting that keeps you out of the instant graduations.</p><h3>Which LetsBonk trigger fits you?</h3><p>If you are not sure where to start, three questions decide most of it.</p><figure><img alt="Decision tree for choosing a LetsBonk sniper trigger: if you know creators with a record, use creator watch with a bigger tip and dev sell on; if you want many small trades, use fresh launch with small size, strict filters, stop loss and a timer; if you prefer proof of demand, use migration with a 50 SOL minimum liquidity, tighter slippage and a trailing stop; otherwise copy trade a trader you trust. All four use a separate sniper wallet and exits on every sniper." src="https://cdn-images-1.medium.com/max/1024/1*IOOYuW5dvebytCTZejyDAw.png" /></figure><ul><li><strong>Do you know creators with a good record?</strong> If yes, start with creator watch. It fires rarely, but each shot has better odds than average.</li><li><strong>Do you want to trade often, with small size?</strong> If yes, use fresh launch with strict filters and a timer. Expect many small losses and a few wins.</li><li><strong>Do you prefer to wait for proof of demand?</strong> If yes, use migration with a liquidity floor. You give up the early price for a coin that has already shown buyers.</li><li><strong>Do you follow a trader you trust?</strong> Copy trade works, but always remember you buy after them, at a higher price on the curve.</li></ul><p>Many people end up with a mix: two or three creator watch snipers, one small fresh launch sniper, and one migration sniper. Each has its own size and exits, and none of them depends on you being awake.</p><h3>Troubleshooting a LetsBonk sniper</h3><p>Here is what usually goes wrong, why, and what to change.</p><ul><li><strong>Creator watch never fires.</strong> The wallet stopped launching, or it launched on another venue you did not tick. Check the wallet’s recent history and tick every venue it might use.</li><li><strong>Buys land at 2x or 3x the first price.</strong> You are too slow behind other buyers. Raise the tip on the snipers where landing first matters most, usually creator watch.</li><li><strong>Buys fail with a slippage error.</strong> Others landed first and the price moved past your cap. Raise the slippage cap a little, or accept that you will miss the most crowded launches.</li><li><strong>Migration buy into an almost empty pool.</strong> You hit an instant graduation with a tiny target. Set a minimum liquidity.</li><li><strong>Sells fail on a falling coin.</strong> The sell slippage is too tight. Give sells more room than buys.</li><li><strong>Lots of small losses on fresh launches.</strong> That is normal for blind snipes. Check that the timer and stop loss are on, the size is small, and the blocklist catches the obvious copycats.</li><li><strong>The sniper “missed” a coin that pumped.</strong> Check whether a filter skipped it. Sometimes that is the filter doing its job: many coins that pump also had a dev holding a big share, and most of those end badly.</li></ul><h3>Choosing a LetsBonk sniper bot</h3><p>There are three kinds of tools people use to snipe LetsBonk coins. Each has a place.</p><h4>Telegram trading bots</h4><p>These are built around buying from a chat. You paste a contract address, pick a size, and the bot buys. Many have sniping features too. They are easy to start with, and most take a share of each trade. They work best for manual trading. They are less suited to running many rules on their own while you are away.</p><h4>Self-hosted bots from GitHub</h4><p>Free code that you run on your own server. You get full control and pay no fee to anyone but your RPC and the network. You also take on every risk of running a stranger’s program that holds your private key. If you go this way, read our checklist on <a href="https://medium.com/@dexsnipecom/raydium-sniper-bot-on-github-9-checks-before-you-run-one-2566b90171e1">Raydium sniper bots on GitHub</a> before you run anything. Also check that the bot actually watches the LaunchLab program. Many older Raydium bots only watch the old AMM v4 program, and they will never see a LetsBonk launch.</p><h4>Hosted rule-based snipers</h4><p>You set the rules in a web panel and the bot runs them on a server. No code, no server to keep alive. The thing to check is how the service handles your key and your money. A good one trades from a separate wallet that you control and fund, so a mistake costs only what is in that wallet.</p><h4>Questions to ask any LetsBonk sniper bot</h4><ol><li>Does it support creator watch on LetsBonk, and does it cover other venues where the same creator might launch?</li><li>Does it run on a server all day, or only while a tab or app is open?</li><li>Does it read the LaunchLab curve directly, including the creator, the quote pair and the graduation target?</li><li>Can it skip bundled launches and cap the dev’s holding at launch?</li><li>Does the migration sniper have a minimum liquidity, so instant graduations do not trigger it?</li><li>How many exits does it have? Take profit steps, stop loss, trailing stop, timer and dev sell are the ones that matter.</li><li>Does it take a cut of each trade, or charge a flat price?</li><li>Does it trade from a separate wallet you control?</li></ol><p>Those questions are the reason we built DexSnipe, our own Solana sniper bot. LetsBonk is its own venue there, with fresh launch, migration, creator watch and copy trade triggers. Every sniper has dev and top 10 holder caps, a bundled launch check, mint and freeze authority checks, a transfer tax cap and a word blocklist, plus up to four take profit steps, a stop loss, a trailing stop, a timer and a dev sell exit, all running on the server after you close the tab. It is a flat plan from 1 SOL for 7 days, with no cut of your trades. You add the key of a separate sniper wallet that you fund yourself.</p><h3>Common LetsBonk sniping mistakes</h3><ol><li><strong>Using one sniper for every moment.</strong> A fresh launch setting with a 20 minute timer will sell a creator watch coin far too early. A creator watch tip on a fresh launch sniper burns SOL on coins you did not choose. Make one sniper per job.</li><li><strong>Following mass launchers.</strong> A wallet with thousands of launches and a handful of graduations is not a signal. It is noise with a long history.</li><li><strong>No liquidity floor on migrations.</strong> This is how you buy into a pool that holds almost nothing.</li><li><strong>Ignoring the quote pair.</strong> A coin paired with another meme coin moves with that coin too. If you only want SOL pairs, filter for them.</li><li><strong>Forgetting the round trip fee.</strong> At about 2% for a buy and a sell on the curve, plus network costs, small take profit targets shrink fast.</li><li><strong>Skipping exits because you trust the creator.</strong> A good creator’s coins still fail most of the time. The exits do the work; the creator only improves the odds.</li><li><strong>Running the bot from your main wallet.</strong> One bad setting, one bad coin, or one bad bot, and the damage is not limited to your trading money.</li><li><strong>Changing five settings at once.</strong> Then you never learn which one helped. Change one, watch ten trades, then decide.</li></ol><h3>Frequently asked questions</h3><h4>What is the best sniper bot for LetsBonk?</h4><p>The best one is the one that fits the moment you want to trade. For creator watch, you need a bot that is armed all day and covers every venue a creator might use. For migration snipes, you need a liquidity floor. For fresh launches, you need fast landing and strong filters. Ask the eight questions above and pick the bot that answers yes to the ones you need.</p><h4>Is LetsBonk the same as Bonk.fun?</h4><p>Yes, you will see both names for the same launchpad. It runs on Raydium LaunchLab, and each of its coins is a LaunchLab curve that records LetsBonk as its platform.</p><h4>How many SOL does a LetsBonk coin need to graduate?</h4><p>A standard SOL-paired LetsBonk curve graduates at 85 SOL. Other quote pairs have targets in their own coin, and LaunchLab lets platforms set other targets, so check the target stored on the curve.</p><h4>Where does a LetsBonk coin go after it graduates?</h4><p>To a Raydium CPMM pool, created automatically with the SOL and tokens left on the curve. The token’s mint authority is revoked at the same time, so no new tokens can be minted.</p><h4>Can the creator pull the liquidity after graduation?</h4><p>For migrations after Raydium’s August 2026 upgrade, Raydium’s docs say the platform and creator share of the LP is locked and the rest is burned. That liquidity cannot be withdrawn. The creator can still sell the tokens they hold, which is why a dev sell exit matters.</p><h4>How often do LetsBonk coins graduate?</h4><p>When we counted 1,150,183 launches on LetsBonk’s three main platform accounts, about 1.1% had graduated. Some creators do much better than that, and mass launchers do worse.</p><h4>What is a migration sniper?</h4><p>A sniper that buys a coin the moment it graduates from the curve and its Raydium pool opens. It skips the curve stage and buys a coin that has already shown 85 SOL of demand.</p><h4>Do I need a Jito tip to snipe LetsBonk?</h4><p>For block zero snipes, it helps a lot. The tip and the priority fee decide where your buy lands in the block, and on a curve the order of buys decides your price. For migration snipes, a smaller tip is usually enough.</p><h4>Is BONKbot a LetsBonk sniper bot?</h4><p>BONKbot is a Telegram trading bot from the same BONK community, but it is a separate product. Like most Telegram bots, it is built around trading from a chat and takes a share of each trade.</p><h4>Can I lose money with a LetsBonk sniper bot?</h4><p>Yes. Most new coins go to zero, and a sniper does not change that. It improves your entry and makes you stick to your exits. Only trade money you can afford to lose.</p><h3>How we got the numbers</h3><p>The launch, creator and graduation counts come from our own read of Raydium’s LaunchLab program in November 2026. We listed every curve account created through LetsBonk’s three main platform accounts, 1,150,183 in total, and read each curve’s creator, status and quote reserve through a Solana RPC node. A curve counted as graduated when its status showed migration to a Raydium pool. For the instant graduations, we read the transaction history of every graduated curve and flagged those with exactly two transactions. The creator sample was 220 creators drawn at random from those with between 2 and 49 launches and at least one graduation. The queue prices use the reserves of a live standard SOL curve and constant product math, before fees. We compared LetsBonk with other launchpads in more depth in <a href="https://medium.com/@dexsnipecom/solana-launchpads-compared-the-hidden-rules-of-letsbonk-jupiter-studio-bags-boop-believe-and-87cef7c5a86d">the hidden rules of Solana launchpads</a>.</p><h3>Sources and further reading</h3><ul><li>Raydium docs, LaunchLab overview (lifecycle, graduation steps, LP after the August 2026 upgrade): <a href="https://docs.raydium.io/products/launchlab/overview">docs.raydium.io/products/launchlab/overview</a></li><li>Raydium docs, LaunchLab bonding curve (curve types, fee fields, worked fee example): <a href="https://docs.raydium.io/products/launchlab/bonding-curve">docs.raydium.io/products/launchlab/bonding-curve</a></li><li>Raydium docs, LaunchLab and CPMM fee reference: <a href="https://docs.raydium.io/products/launchlab/tips-and-gotchas/launchlab-cpmm-fee-reference">docs.raydium.io</a></li><li>Raydium docs, buying a LaunchLab token (user flow, slippage, risks): <a href="https://docs.raydium.io/user-flows/buying-a-launchlab-token">docs.raydium.io/user-flows/buying-a-launchlab-token</a></li><li>Raydium LaunchLab program on Solscan: <a href="https://solscan.io/account/LanMV9sAd7wArD4vJFi2qDdfnVhFxYSUg6eADduJ3uj">solscan.io</a></li><li>Main LetsBonk platform account on Solscan: <a href="https://solscan.io/account/FfYek5vEz23cMkWsdJwG2oa6EphsvXSHrGpdALN4g6W1">solscan.io</a></li><li>Cointelegraph, LetsBonk passes Pump.fun in daily revenue (launch date and LaunchLab partnership): <a href="https://cointelegraph.com/news/letsbonk-surpasses-pumpfun-daily-revenue-defillama">cointelegraph.com</a></li><li>Decrypt, BONK launchpad flips Pump.fun for Solana tokens created: <a href="https://decrypt.co/328822/bonk-meme-coin-launchpad-flips-pump-fun-for-solana-tokens-created">decrypt.co</a></li><li>DefiLlama, Bonk.fun launchpad fees: <a href="https://defillama.com/fees/bonk.fun-launchpad">defillama.com</a></li><li>Jito docs, tips and bundles on Solana: <a href="https://docs.jito.wtf/lowlatencytxnsend/">docs.jito.wtf</a></li></ul><p>Docs were checked on November 13, 2026. Nothing here is financial advice. Meme coins are very risky, and most new tokens go to zero. Only trade with money you can afford to lose.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0a785acffa8c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Raydium Sniper Bot on GitHub: 9 Checks Before You Run One]]></title>
            <link>https://medium.com/@dexsnipecom/raydium-sniper-bot-on-github-9-checks-before-you-run-one-2566b90171e1?source=rss-2654b58e3d1e------2</link>
            <guid isPermaLink="false">https://medium.com/p/2566b90171e1</guid>
            <category><![CDATA[trading-bot]]></category>
            <category><![CDATA[cryptosecurity]]></category>
            <category><![CDATA[github]]></category>
            <category><![CDATA[solanas]]></category>
            <category><![CDATA[raydium]]></category>
            <dc:creator><![CDATA[DexSnipe]]></dc:creator>
            <pubDate>Thu, 12 Nov 2026 12:15:36 GMT</pubDate>
            <atom:updated>2026-11-12T12:15:36.015Z</atom:updated>
            <content:encoded><![CDATA[<h3>Free sniper bots on GitHub can be a great way to learn. Some have also emptied wallets. Here is how to read one before you give it a private key.</h3><p>Search for a Raydium sniper bot and GitHub shows up near the top. The pitch is simple: the code is free, you can read it, and you run it yourself. That is all true. But a sniper bot is a program that holds a private key and signs transactions with it while you sleep, so the code you download decides what happens to your money. Most people clone the repo, paste a key into a .env file and press start. This guide is about the ten minutes you should spend before that. We will walk through what a Raydium sniper bot actually does, what went wrong for people who ran the wrong code, and nine checks you can do yourself, even if you do not write code for a living.</p><figure><a href="https://www.dexsnipe.com/"><img alt="DexSnipe Raydium sniper bot panel illustration: a web panel with a Raydium sniper turned on. The activity log shows pools skipped because the LP was held by the opener, freeze authority was on, the pool was under the minimum SOL, a 5% transfer fee or mint authority was on, and one CPMM pool with burned LP that passed, was bought for 0.05 SOL and hit a +50% take profit." src="https://cdn-images-1.medium.com/max/1024/1*MSttOTf10VAtcsp6rOsnjA.png" /></a><figcaption>DexSnipe is our own Solana sniper bot service.</figcaption></figure><h3>The short version</h3><ul><li><strong>Check which Raydium programs the bot watches.</strong> Raydium now has four places where new tokens can trade: AMM v4, CPMM, CLMM and LaunchLab. Raydium’s own docs say new standard pools go to CPMM and CLMM, while AMM v4 is kept for existing pools. Many older GitHub bots only watch AMM v4.</li><li><strong>Check when the code last changed.</strong> The most starred open source Solana trading bot we looked at has over 2,300 stars, and its last code push was in August 2024.</li><li><strong>Follow your private key through the code.</strong> Search for the name of the key variable and look at every line that touches it. It should only be used to sign.</li><li><strong>Read the dependency list.</strong> In 2025, a fake Solana bot on GitHub hid its key stealer inside an npm package with a harmless sounding name.</li><li><strong>Install with the lockfile.</strong> In December 2024, two versions of the official Solana JavaScript library were published with a backdoor. A loose version range plus a fresh install could have pulled them in.</li><li><strong>First run on a burner wallet with a small amount.</strong> Whether you run a GitHub bot or a hosted one, the sniper wallet should only hold what you plan to trade.</li></ul><h3>What does a Raydium sniper bot actually do?</h3><p>Before you can judge any bot’s code, it helps to know what the code is supposed to do. Every sniper bot, on any DEX, runs the same loop:</p><ol><li><strong>Listen.</strong> The bot opens a connection to a Solana RPC node and listens for new pools on the programs it cares about.</li><li><strong>Read the pool.</strong> When a new pool shows up, it reads the pool account: which token, how much liquidity, who opened it.</li><li><strong>Filter.</strong> It runs your rules. Is the LP burned or locked? Can the creator still mint more tokens or freeze your account? Is there enough SOL in the pool?</li><li><strong>Buy.</strong> It builds a swap with your size, your slippage limit and your priority fee, signs it with your key and sends it.</li><li><strong>Exit.</strong> It watches the price and sells on your take profit, stop loss, trailing stop or timer.</li></ol><p>Steps one and four are where most GitHub bots differ, and they are also where most of the risk sits. Step one decides whether the bot even sees the pools you want. Step four is the only place your private key should be used.</p><h4>The four Raydium programs</h4><p>“Raydium” is not one program. It is a family of on-chain programs, each with its own address. A bot that listens to the wrong one will sit there and see nothing, or see only old style pools. Here are the four that matter for sniping, with the addresses from Raydium’s official program address page:</p><ul><li><strong>AMM v4</strong>, address starting 675kPX9M. The classic Raydium pool, linked to an OpenBook market. Raydium&#39;s docs say it is kept for existing pools, and that the Raydium app and SDK send new pool creation to CPMM and CLMM.</li><li><strong>CPMM</strong>, address starting CPMMoo8L. The standard constant product pool Raydium uses now. No OpenBook market needed, and it supports Token-2022 tokens.</li><li><strong>CLMM</strong>, address starting CAMMCzo5. Concentrated liquidity pools. Less common for brand new meme coins, more common for larger tokens.</li><li><strong>LaunchLab</strong>, address starting LanMV9sA. Raydium&#39;s launchpad, with a bonding curve that later moves the token into a regular pool. We covered how it differs from other launchpads in our post on <a href="https://medium.com/@dexsnipecom/solana-launchpads-compared-the-hidden-rules-of-letsbonk-jupiter-studio-bags-boop-believe-and-87cef7c5a86d">the hidden rules of Solana launchpads</a>.</li></ul><figure><img alt="Diagram of the four Raydium programs a sniper bot can watch: AMM v4 (address starting 675kPX9M), the classic pool with an OpenBook market kept for existing pools; CPMM (CPMMoo8L), the standard pool for new tokens with Token-2022 support; CLMM (CAMMCzo5), concentrated liquidity for larger tokens; and LaunchLab (LanMV9sA), Raydium’s launchpad. A bot that listens only to AMM v4 can miss new CPMM pools." src="https://cdn-images-1.medium.com/max/1024/1*eBwYfA5uqS7ldTeKQc3pYg.png" /></figure><p>Always copy the full addresses from Raydium’s own docs, not from a blog post, a Telegram message or this article. Raydium’s docs call that page the source of truth, and it is the first thing you will compare the bot’s code against in check 3.</p><h3>Why people start with a GitHub bot</h3><p>There are good reasons to look at open source sniper bots, and it is worth saying them clearly.</p><ul><li><strong>It costs nothing to try.</strong> You pay for an RPC plan and a server, but not for the bot itself.</li><li><strong>You can see what it does.</strong> With a closed tool you take the vendor’s word for it. With open code you can, in theory, read every line.</li><li><strong>You learn a lot.</strong> Reading a working sniper is one of the fastest ways to understand how Solana transactions, priority fees and pool accounts fit together.</li><li><strong>You can change it.</strong> If you want a filter nobody offers, you can write it.</li></ul><p>The catch is in the words “in theory”. Most people who download a sniper bot do not read it. They read the README, fill in the .env file and run it. And attackers know this, which is why sniper bots have become one of the most popular disguises for wallet stealing code.</p><h3>What went wrong for people who ran the wrong code</h3><p>These are three real cases, all public, all with sources at the end of this post. None of them needed the victim to do anything unusual. They just ran code.</p><h4>Case 1: a fake Solana bot with a hidden dependency (2025)</h4><p>In July 2025, the security firm SlowMist looked into a GitHub project called solana-pumpfun-bot after a user lost funds running it. The bot itself looked normal. The theft sat in one of its dependencies, a package called crypto-layout-utils. npm has since removed it. Its code was heavily obfuscated, and when it ran it searched the computer for wallet and private key files and sent them to a server run by the attacker. SlowMist also found other GitHub accounts forking the project and spreading versions with a similar package, which made the repo look used and trusted. The stolen funds were moved through an exchange.</p><p>The lesson: the file you need to read is not only the bot. It is the list of packages the bot pulls in.</p><h4>Case 2: a backdoor in the official Solana library (December 2024)</h4><p>On December 2, 2024, two versions of @solana/web3.js, the main JavaScript library for Solana, were published to npm with a backdoor. Versions 1.95.6 and 1.95.7 added a function called addToQueue that sent private keys to the attacker when the library touched them. The cause was a stolen publish account, not a bug in Solana. The bad versions were pulled within hours and 1.95.8 removed the code. Security researchers noted that the people most at risk were bots and backend scripts that handle private keys directly, which is exactly what a sniper bot is.</p><p>The lesson: even a perfectly honest bot can pull in bad code, if you install it in a way that grabs the newest version of everything.</p><h4>Case 3: fake stars on GitHub</h4><p>A team of researchers from Carnegie Mellon, Socket and North Carolina State built a tool called StarScout and ran it on GitHub data from July 2019 to December 2024. They found about 6 million suspected fake stars across more than 18,000 repositories, and said the trend grew sharply in 2024. Many of the repos with fake stars were short lived ones pushing malware, with crypto bots among the common disguises.</p><p>The lesson: a star count is a marketing number. It tells you nothing about whether the code is safe.</p><figure><img alt="Diagram of six places a bad GitHub sniper bot repo hides key theft: look-alike helper packages in package.json dependencies, preinstall or postinstall scripts, the private key passed to anything other than signing, Telegram or Discord webhooks and unknown URLs, unreadable obfuscated code with eval, and README links to password protected zips or exe files. Notes the 2025 fake Solana bot that hid its stealer in a dependency and the December 2024 @solana/web3.js 1.95.6 and 1.95.7 backdoor." src="https://cdn-images-1.medium.com/max/1024/1*wblhApuseOU7oAqVatK6Dg.png" /></figure><h3>The 9 checks</h3><p>None of these checks need you to be a professional developer. You need a browser, a text editor and the search function. If a repo fails one of the first eight checks, do not run it. The ninth check is how you run it if it passes.</p><h4>Check 1: who made it, and are the stars real?</h4><p>Open the profile of the account that owns the repo. A real developer usually has a history: other projects, commits spread over months or years, issues they opened or answered somewhere else. A throwaway account usually has one or two repos, all created in the same week, and no activity outside them.</p><p>Then look at who starred the repo. Click the star count and open a handful of the accounts. If most of them have no avatar, no repos and joined recently, the stars are probably bought. The same goes for forks. In the SlowMist case, the forks were part of the trick.</p><p>A good sign: the issues tab is open and has real questions from real users, with real answers. A bad sign: issues are turned off, or every issue is a thank you message.</p><h4>Check 2: when was the code last changed?</h4><p>Click the commits link and look at the date of the last change to the actual code, not just the README. Solana and Raydium both change. Raydium’s own docs list a CPMM program upgrade and a CLMM program upgrade in September 2026 alone. A sniper bot that has not been touched in a year or more may still run, but it may be reading account layouts or calling instructions in ways that no longer match.</p><p>Here is a real example. The most starred open source Solana trading bot we found, warp-id/solana-trading-bot, had 2,346 stars and 971 forks when we checked on November 12, 2026. Its last code push was on August 10, 2024. To be fair to the author, the README is honest and labels it a beta. But it tells you something about how popular and how current are two different things.</p><h4>Check 3: which Raydium programs does it watch?</h4><p>This is the check most people skip, and it is the one that most often explains “my bot runs but never buys anything”.</p><p>Open the file that sets up the listener. It is often called something like listener, monitor, watcher or stream. Look for the program address or a constant name like AmmV4, CPMM, CLMM or LaunchLab. Then compare it to the list above.</p><p>In the warp-id bot, the listener subscribes to the AMM v4 program and to OpenBook markets. That made sense in 2024, when most new Raydium pools were AMM v4 pools with an OpenBook market. Today, Raydium’s docs say new standard pools go to CPMM. A bot that only watches AMM v4 will miss those, no matter how fast it is.</p><p>Ask yourself three questions here:</p><ul><li>Does it watch CPMM?</li><li>Does it handle LaunchLab tokens, both on the curve and after they move to a pool?</li><li>If it watches several programs, does it read each pool layout correctly, or does it assume they are all the same?</li></ul><h4>Check 4: read package.json line by line</h4><p>In a JavaScript or TypeScript bot, open package.json and read the dependencies list. It is usually short, between ten and thirty lines. For each package, you want to know two things: what it is, and whether it is the real one.</p><p>Names you should expect in a Solana sniper bot: @solana/web3.js, @solana/spl-token, @raydium-io/raydium-sdk or raydium-sdk-v2, bs58, bn.js, dotenv, some logger. These are the normal building blocks.</p><p>Names that should make you stop: anything that sounds like a helper for a normal package but is not the normal package. A “bs58-encrypt-utils” next to bs58. A “crypto-layout-utils” next to the official buffer layout packages. In the SlowMist case, that is exactly what the stealer looked like.</p><p>For any name you do not know, open its page on npmjs.com and check three things: how many weekly downloads it has, who publishes it, and when the first version came out. A package with almost no downloads, first published a few weeks ago, by an account with no other packages, has no business being in a bot that touches your key.</p><p>Also look at the scripts section. A script named preinstall, install or postinstall runs code the moment you install the packages, before you have run the bot at all. Most sniper bots do not need one. If there is one, read exactly what it runs.</p><h4>Check 5: is there a lockfile, and do you use it?</h4><p>Next to package.json there should be a package-lock.json, yarn.lock or pnpm-lock.yaml. This file pins the exact version of every package, including the packages your packages depend on.</p><p>Why it matters: in package.json, versions are often written as ranges, like ^1.89.1. That caret means “this version or any newer 1.x”. If you install without a lockfile, npm takes the newest version that fits. If you had run a fresh install of a bot with “@solana/web3.js”: “¹.89.1” during the few hours on December 2, 2024 when the backdoored versions were live, you could have received one of them.</p><p>The warp-id bot does ship a lockfile, and it pins web3.js at 1.89.1. That is the good pattern. To use it, install with npm ci rather than npm install. The npm docs describe npm ci as the command for clean, repeatable installs: it installs exactly what the lockfile says and stops with an error if the lockfile and package.json do not match. Never delete the lockfile to &quot;fix&quot; an install error.</p><h4>Check 6: follow your private key through the code</h4><p>This is the most important check, and it is easier than it sounds.</p><p>Find the name the bot uses for your key. Look in the README or the .env.example file. It is usually PRIVATE_KEY, WALLET_PRIVATE_KEY, SECRET_KEY or similar. Now use the search in your editor, or GitHub’s own code search, to find every place that name appears.</p><p>In an honest bot, you will see a short path:</p><ol><li>The key is read from the environment.</li><li>It is decoded, usually from base58, into a Keypair object.</li><li>That Keypair is used to sign transactions.</li></ol><p>That is it. Then follow the Keypair variable the same way. It should only show up where transactions are signed and where the public address is read.</p><p>Red flags: the key or the Keypair is passed into a function whose name has nothing to do with signing. It is written to a file. It is joined into a string. It is added to a request body, a header or a URL. It is logged, even “only in debug mode”. Any of those is enough to stop.</p><p>A tip: also search for the words secretKey and fromSecretKey. Some bots read the key under one name and pass it around under another.</p><h4>Check 7: list every place the bot sends data</h4><p>A sniper bot needs to talk to a few places: your RPC endpoint, maybe a Jito block engine for bundles, maybe a price API. That list should be short and you should recognise all of it.</p><p>Search the code for http, https, wss, fetch, axios, got, request and WebSocket. Write down every address you find. Then search for the words telegram, discord, webhook and api.telegram.org. A Telegram or Discord webhook inside a sniper bot is the most common way stolen keys are sent out, because it is free and easy to set up.</p><p>Some bots send transactions through the author’s own service. That is not bad by itself, as long as it is clearly disclosed. The warp-id README, for example, explains that its optional “warp” executor sends the signed transaction to a hosted service, says the payload does not contain your private key, and says the option is off by default. That is the kind of plain disclosure you want. What you do not want is an outbound call nobody mentions.</p><h4>Check 8: look for code you cannot read</h4><p>Open the source files and scroll. Normal code has normal names and normal line lengths. Be very careful with:</p><ul><li>One huge line that runs off the right side of the screen.</li><li>Long blocks of random looking letters and numbers, often base64 or hex.</li><li>Variable names like _0x3f2a, _0x1b9c, again and again.</li><li>Calls to eval, Function(), or anything that builds code from a string and runs it.</li><li>A file that is clearly minified, inside a project that is otherwise plain TypeScript.</li></ul><p>In the SlowMist case, the stealer was obfuscated with a common JavaScript obfuscation service. There is no honest reason for a free, open source sniper bot to hide part of its own code from you. The whole point of open source is that you can read it.</p><p>The same goes for anything that asks you to download a password protected zip, run a .exe, or “get the full version” through a link in the README. Close the tab.</p><h4>Check 9: first run on a burner wallet</h4><p>If the repo passed checks 1 to 8, you still do not run it on your main wallet. Here is the safe way to do a first run:</p><ol><li><strong>Create a new wallet just for the bot.</strong> A fresh keypair, never used for anything else, never holding NFTs or long term coins.</li><li><strong>Send it a small amount.</strong> Enough for a few trades and fees, nothing you would miss.</li><li><strong>Run it on a separate machine if you can.</strong> A cheap VPS or a virtual machine is better than your daily laptop, because the 2025 stealer searched the whole computer for wallet files, not just the bot folder.</li><li><strong>Watch it for a day.</strong> Check the wallet on a block explorer. Every transaction should be one you expect: buys, sells, fees. An unexpected transfer out means stop, and move what is left.</li></ol><p>Keep this habit even after the bot proves itself. The sniper wallet should only ever hold what you are willing to trade with. Top it up from your main wallet when you need to; never the other way round.</p><figure><img alt="Checklist of 9 checks before you run a Raydium sniper bot from GitHub: who made it and real stargazers, last code change, which Raydium programs it watches (CPMM and LaunchLab, not only AMM v4), known packages and no install scripts in package.json, a lockfile installed with npm ci, the private key path limited to read, decode and sign, named outbound calls only, readable code with no obfuscation or eval, and a first run on a burner wallet with a small amount on a separate machine." src="https://cdn-images-1.medium.com/max/1024/1*otX_DJ170kk8ZeMZ38XRMQ.png" /></figure><h3>If the repo passes: setting it up properly</h3><p>A safe bot can still lose money if it is set up badly. Here is what to get right, in order.</p><h4>Your RPC connection</h4><p>Free public RPC endpoints are fine for reading a balance. They are not fine for sniping. They limit how many requests you can make, and many bots warn in their README that some options should not be used with a public RPC. You will want a paid RPC plan with a websocket endpoint at the very least. The Solana docs describe the standard websocket subscriptions, like logsSubscribe and programSubscribe, that most simple bots use. Faster setups use a streaming feed such as Yellowstone gRPC, which most RPC providers offer on higher plans. Pick the server location close to your RPC provider, not close to you.</p><h4>Your buy settings</h4><p>Three numbers decide whether your buy lands and what it costs: slippage, priority fee and, if you use Jito, the tip. They work against each other. Higher slippage and a bigger fee make the buy more likely to land, and more likely to land at a bad price. We tested how these play out on real launches in our post on <a href="https://medium.com/@dexsnipecom/best-sniper-bot-settings-for-pumpfun-slippage-priority-fee-and-tip-from-807-launches-76409e7a5040">the best sniper bot settings for slippage, priority fee and tip</a>. The same logic applies on Raydium.</p><p>Two Raydium specific notes. First, make sure the bot sets a compute unit limit and price that fit the program it is calling; a CLMM swap uses more compute than a CPMM swap. Second, if the bot supports both SOL and USDC pools, decide which you want and set it explicitly.</p><h4>Your filters</h4><p>On Raydium, the creator chose how to open the pool, and that gives them more room to hurt you than on a fixed launchpad. These are the filters worth having on from day one:</p><ul><li><strong>LP burned or locked.</strong> If the person who opened the pool still holds the LP tokens, they can pull the liquidity. A bot should check this before buying. Some GitHub bots have it as an option that is off by default, so check your config.</li><li><strong>Mint authority off.</strong> If it is still on, the creator can print more tokens.</li><li><strong>Freeze authority off.</strong> If it is still on, the creator can freeze your token account and you cannot sell.</li><li><strong>Minimum pool size.</strong> Very small pools move a lot on every buy, including yours.</li><li><strong>Token-2022 transfer fee.</strong> CPMM supports Token-2022 tokens, and some of those charge a fee on every transfer. Raydium’s docs also show that LaunchLab can create a token with a transfer fee of up to 5%. A bot that does not check it can buy a token that takes a cut of every sell.</li></ul><p>We wrote a full piece on which filters actually helped on new launches in <a href="https://medium.com/@dexsnipecom/pump-fun-sniper-bot-filters-we-checked-807-new-coins-to-see-which-rules-work-5c7a932f9e9e">Pump.fun sniper bot filters</a>. The creator history and creator buy rules there carry over well to LaunchLab tokens.</p><h4>Your exits</h4><p>Most GitHub sniper bots focus on the buy and treat the sell as an afterthought: one take profit, one stop loss, maybe a timer. That is not enough for new pools, where the first minutes decide most outcomes. At the least you want a take profit that sells part of the bag, a stop loss, and a time limit so the bot does not hold a dead token forever. Our first post, on <a href="https://medium.com/@dexsnipecom/how-to-snipe-meme-coins-on-solana-what-2-023-replayed-trades-say-about-timing-fees-and-exits-524c50d4e00e">how to snipe meme coins on Solana</a>, goes into how timing and exits changed the result of the same trades.</p><h3>What running your own bot really takes</h3><p>Here is the part the README rarely mentions. A self hosted sniper bot is a small server you now look after.</p><ul><li><strong>A server that is always on.</strong> Your laptop going to sleep means missed pools and open positions with no stop loss.</li><li><strong>A paid RPC plan.</strong> Usually the biggest monthly cost, and the one that decides your speed.</li><li><strong>Updates.</strong> When Raydium upgrades a program or Solana changes how fees work, someone has to update the code. If the repo is abandoned, that someone is you.</li><li><strong>Security, forever.</strong> Every time you pull new code or update packages, checks 4 to 8 start again.</li><li><strong>Your time.</strong> Logs to read, crashes to restart, settings to tune.</li></ul><p>For some people this is the fun part, and if you enjoy it, a well chosen GitHub bot is a great teacher. For others, it is the reason they never get past the setup.</p><figure><img alt="Decision tree for choosing a GitHub sniper bot or a hosted sniper bot: if you cannot read TypeScript well enough to follow a variable through a few files, learn first or go hosted; if you can but will not look after a server with an RPC plan, updates, restarts and logs, use a hosted sniper bot; if both are yes, run a GitHub bot after the 9 checks, install with npm ci and start on a burner wallet. Either way the sniper wallet only holds what you plan to trade." src="https://cdn-images-1.medium.com/max/1024/1*vResDY_6K2CBA3ab6BrvlA.png" /></figure><h3>When a hosted Raydium sniper bot makes more sense</h3><p>A hosted bot runs the server, the RPC connection and the updates for you. You set your rules in a web panel and the bot runs them. The trade off is that you trust the service instead of auditing the code yourself, so the same wallet rule applies: use a separate sniper wallet that only holds what you plan to trade.</p><p>That is how we built <a href="https://www.dexsnipe.com/">DexSnipe</a>, our own Solana sniper bot. Raydium is one of the venues you can tick in a sniper, next to Pump.fun, PumpSwap, LaunchLab, Meteora and others. Each sniper can watch for new pools, a creator wallet you choose, or a wallet to copy. Before every buy it runs the checks from this post: LP burned or locked, mint and freeze authority, minimum liquidity in SOL, a transfer tax cap, dev and top 10 holder caps, and a word blocklist for names and tickers. After the buy, you pick from five exits. You can build a sniper in the panel for free; a flat plan from 1 SOL for 7 days switches it on, with no cut taken from your trades.</p><p>Whichever way you go, the questions in this post are the right ones to ask. A good hosted service should be able to answer them as clearly as a good repo.</p><h3>Common mistakes with Raydium sniper bots</h3><ul><li><strong>Trusting the star count.</strong> Stars and forks can be bought. Read the code and the people behind it instead.</li><li><strong>Running npm install on an old repo with no lockfile.</strong> You get whatever is newest today, which is not what the author tested.</li><li><strong>Using your main wallet.</strong> One bad dependency and everything in it is gone.</li><li><strong>Watching only AMM v4.</strong> The bot runs, the logs look healthy, and it never sees a new CPMM pool.</li><li><strong>Leaving the LP check off.</strong> On Raydium, the person who opened the pool can often still pull it. This is the filter that matters most.</li><li><strong>Testing on a free RPC.</strong> The bot looks slow and broken, when the real problem is the rate limit.</li><li><strong>No time limit on positions.</strong> A token that stops trading never hits your stop loss. Without a timer, the bot holds it forever.</li><li><strong>Pasting a seed phrase.</strong> No sniper bot needs your 12 or 24 words. A bot that asks for a seed phrase instead of a single wallet key is asking for every wallet you ever made from it.</li></ul><h3>Questions people ask about Raydium sniper bots</h3><h4>Is it safe to use a Raydium sniper bot from GitHub?</h4><p>It can be, if you read the code first and run it on a separate wallet with a small amount. The risk is not GitHub itself. It is running code you have not checked with a key that can move your money.</p><h4>Which Raydium program should a sniper bot watch?</h4><p>For new pools today, CPMM above all, plus LaunchLab if you want tokens from Raydium’s launchpad. AMM v4 is kept for existing pools. CLMM is used more for larger tokens. Check Raydium’s program address page for the current list.</p><h4>Why does my Raydium sniper bot run but never buy?</h4><p>The most common reasons: it is listening to a program where few new pools appear, its filters are too strict, the RPC is dropping websocket events, or its buys are failing because slippage or priority fee are too low. Check the logs for “new pool” lines first. If there are none, it is the listener.</p><h4>How do I know if a sniper bot repo is malware?</h4><p>Look for unknown npm packages with few downloads, install scripts that run on their own, code you cannot read, outbound calls to Telegram or Discord, and any place your private key is used for anything other than signing. Any one of these is a reason not to run it.</p><h4>Should I give a sniper bot my private key or my seed phrase?</h4><p>A bot that trades for you needs a private key to sign, whether it runs on your machine or on a server. Give it the key of a separate wallet made only for the bot. Never give any bot a seed phrase.</p><h4>Do I need to know how to code to check a GitHub bot?</h4><p>Not much. Checks 1, 2 and 9 need no code at all. Checks 3 to 8 need you to open files and use search. If you cannot follow the private key through the code after an honest try, that is a good sign the bot is not for you.</p><h4>Is a hosted Raydium sniper bot faster than a GitHub one?</h4><p>It depends on the setup. A self hosted bot on a strong RPC plan and a nearby server can be very fast. A self hosted bot on a free RPC and a home connection will be slow. A hosted bot gives you a fixed setup without having to build it.</p><h3>Sources and further reading</h3><ul><li>Raydium docs, program addresses (the source of truth for program IDs): <a href="https://docs.raydium.io/reference/program-addresses">docs.raydium.io/reference/program-addresses</a></li><li>Raydium docs, versions and migration between pool programs: <a href="https://docs.raydium.io/protocol-overview/versions-and-migration">docs.raydium.io/protocol-overview/versions-and-migration</a></li><li>Raydium official source repositories: <a href="https://github.com/raydium-io">github.com/raydium-io</a></li><li>warp-id/solana-trading-bot, the open source bot used as an example: <a href="https://github.com/warp-id/solana-trading-bot">github.com/warp-id/solana-trading-bot</a></li><li>Cointelegraph on SlowMist’s findings about the fake solana-pumpfun-bot repo: <a href="https://cointelegraph.com/news/solana-trading-bot-github-malware-scam">cointelegraph.com</a></li><li>Socket on the @solana/web3.js supply chain attack: <a href="https://socket.dev/blog/supply-chain-attack-solana-web3-js-library">socket.dev/blog</a></li><li>ReversingLabs on the malware found in the Solana npm library: <a href="https://reversinglabs.com/blog/malware-found-in-solana-npm-library-with-50m-downloads">reversinglabs.com/blog</a></li><li>He et al., “Six Million (Suspected) Fake Stars on GitHub”, arXiv 2412.13459: <a href="https://arxiv.org/abs/2412.13459">arxiv.org/abs/2412.13459</a></li><li>npm docs, npm ci: <a href="https://docs.npmjs.com/cli/v10/commands/npm-ci">docs.npmjs.com</a></li><li>Solana docs, websocket methods: <a href="https://solana.com/docs/rpc/websocket">solana.com/docs/rpc/websocket</a></li></ul><p>Repository details were checked on GitHub on November 12, 2026. Nothing here is financial advice. Meme coins are very risky, and most new tokens go to zero. Only trade with money you can afford to lose.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2566b90171e1" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Pump.fun Sniper Bot Filters: We Checked 807 New Coins to See Which Rules Work]]></title>
            <link>https://medium.com/@dexsnipecom/pump-fun-sniper-bot-filters-we-checked-807-new-coins-to-see-which-rules-work-5c7a932f9e9e?source=rss-2654b58e3d1e------2</link>
            <guid isPermaLink="false">https://medium.com/p/5c7a932f9e9e</guid>
            <category><![CDATA[memecoins]]></category>
            <category><![CDATA[solana-blockchain]]></category>
            <category><![CDATA[cryptocurrency]]></category>
            <category><![CDATA[trading-bot]]></category>
            <category><![CDATA[crypto-trading]]></category>
            <dc:creator><![CDATA[DexSnipe]]></dc:creator>
            <pubDate>Wed, 11 Nov 2026 17:13:07 GMT</pubDate>
            <atom:updated>2026-11-11T17:13:07.098Z</atom:updated>
            <content:encoded><![CDATA[<h3>Most new Pump.fun coins stop trading within an hour. We followed 807 launches for a full day and tested the filters a sniper bot can run before it buys.</h3><p>A Pump.fun sniper bot buys new coins in the first seconds after launch. Speed gets most of the attention, but the bigger question is what the bot buys. Hundreds of coins launch every hour, and a bot with no filters buys all of them. So we took the 807 Pump.fun coins created between 02:05 and 04:05 UTC on November 10, 2026, and checked what happened to each one over the next 24 hours. Then we looked at five things a bot can check at launch and asked a simple question: which of them would have kept you out of the coins that died fast?</p><figure><a href="https://www.dexsnipe.com/"><img alt="DexSnipe dashboard illustration with Pump.fun filters on, skipping coins for a blocklisted creator, an 18% dev holding, 4 bundle buyers, no creator buy and a 41% top 10 share, and a button to open dexsnipe.com" src="https://cdn-images-1.medium.com/max/1024/1*Y60gytn2_TmiV-caJ_oi3Q.png" /></a><figcaption>DexSnipe is our own Solana sniper bot service.</figcaption></figure><h3>The short answer</h3><p>Here is what the data says, before the details.</p><ul><li><strong>Most new coins are over in minutes.</strong> 342 of 807 coins (42%) had their last bonding curve trade within 10 minutes of launch. Only 29% were still trading after one hour. 15 coins graduated, and 7 of those were bought out by the creator or one or two wallets within seconds.</li><li><strong>Creator history was the strongest filter.</strong> 38 wallets launched five or more coins each in the two hours. Together they made 378 coins, 47% of all launches. 53% of their coins were dead within 10 minutes, against 29% for creators who launched only once.</li><li><strong>A creator who buys nothing is a warning sign.</strong> 281 coins had no buy from the creator at launch. 60% of them were dead within 10 minutes. The three busiest creators launched 99 coins this way, and none of those coins was still trading after an hour.</li><li><strong>A big creator buy keeps a coin busy, but busy is not the same as good.</strong> Coins where the creator took 5% or more stayed active longer, but only one of those 102 coins graduated with real buyers.</li><li><strong>No filter finds winners.</strong> Only 8 coins (1%) graduated with real buyers. Filters are good at cutting coins that die fast. They do not pick the next big coin for you.</li></ul><p>The rest of this post walks through each filter, what it caught, and the settings we would start with. If you want to know how fast you need to be, our post on <a href="https://medium.com/@dexsnipecom/solana-sniper-bot-who-really-buys-first-we-traced-411-early-pumpfun-buys-6cc602a342c2">who really buys first on Pump.fun</a> covers that part.</p><h3>What is a sniper bot filter?</h3><p>A filter is a rule the bot checks after it sees a new coin and before it sends the buy. If the coin fails the rule, the bot skips it. Every sniper bot works in four steps:</p><ol><li><strong>Watch.</strong> The bot listens for new create transactions on Pump.fun.</li><li><strong>Filter.</strong> It checks the new coin against your rules.</li><li><strong>Buy.</strong> It sends a buy with your size, slippage limit and fee.</li><li><strong>Exit.</strong> It sells on your take profit, stop loss or timer.</li></ol><p>Step two is where most of your results come from. A bot that buys every coin pays fees on every coin. If 4 in 10 coins stop trading within 10 minutes, a large part of your budget goes into coins that never had a chance. Good filters do two things: they cut the number of bad buys, and they cut the number of trades, so you pay less in fees and tips.</p><p>Some filters can only be checked after the coin has traded for a moment. That matters for a sniper, because waiting costs you price. We come back to this below, because it changes how you should set up your bot.</p><h3>How we collected the data</h3><p>We did not use any third-party dashboard. Everything comes from Solana mainnet.</p><ol><li><strong>The coins.</strong> We pulled every successful create transaction from the Pump.fun program between 02:05 and 04:05 UTC on November 10, 2026. That gave us 807 coins from 343 creator wallets. This is the same set of coins we used in our post on <a href="https://medium.com/@dexsnipecom/best-sniper-bot-settings-for-pumpfun-slippage-priority-fee-and-tip-from-807-launches-76409e7a5040">the best sniper bot settings for Pump.fun</a>.</li><li><strong>The first ten seconds.</strong> For each coin we read every transaction that touched its bonding curve in the first 25 slots, about ten seconds. From the change in SOL on the curve we could tell buys from sells and who sent them.</li><li><strong>The next 24 hours.</strong> On November 11 we asked the chain for the full list of transactions on each bonding curve from launch to 24 hours later. We counted trades in the first hour and noted the time of the last trade.</li><li><strong>The final state.</strong> We read each bonding curve account to see if the coin had graduated.</li><li><strong>The metadata.</strong> We read the name, symbol and metadata link from each create transaction. The link shows where the coin’s metadata file is hosted.</li></ol><p>A few definitions, so the numbers are clear:</p><ul><li><strong>Dead within 10 minutes</strong> means the last trade on the bonding curve came less than 10 minutes after launch, and the coin did not graduate.</li><li><strong>Still trading after one hour</strong> means the curve had at least one trade more than 60 minutes after launch, or the coin graduated with real buyers.</li><li><strong>Graduated with real buyers</strong> means the curve filled and the coin moved to a PumpSwap pool, after hundreds or thousands of trades.</li><li><strong>Bought out</strong> means the curve filled within a handful of trades, by the creator or one or two wallets in the first seconds.</li><li><strong>Creator buy share</strong> is the part of the 1 billion supply the creator bought in the create transaction. We worked it out from the bonding curve formula and the SOL that went in.</li></ul><p>Two limits to keep in mind. First, this is two hours of launches on one day. Patterns can change week to week. Second, a trade count says nothing about who is trading. Some of the busiest curves may be mostly bots trading with each other. We point this out where it matters.</p><h3>What happened to 807 new Pump.fun coins?</h3><p>Before testing filters, it helps to see the baseline. This is what a bot with no filters would have been buying into.</p><figure><img alt="Bar chart of when 807 new Pump.fun coins created on November 10 2026 made their last bonding curve trade: 172 under 1 minute, 170 in 1 to 10 minutes, 235 in 10 to 60 minutes, 64 in 1 to 6 hours and 166 after 6 hours. 42% were done within 10 minutes. 15 graduated, 8 after real trading and 7 bought out in seconds." src="https://cdn-images-1.medium.com/max/1024/1*oqBKl4uxVvvY5tuRzgbMKg.png" /></figure><p>Here is how long the 807 coins kept trading on the bonding curve:</p><ul><li><strong>Under 1 minute:</strong> 172 coins (21%).</li><li><strong>1 to 10 minutes:</strong> 170 coins (21%).</li><li><strong>10 to 60 minutes:</strong> 235 coins (29%).</li><li><strong>1 to 6 hours:</strong> 64 coins (8%).</li><li><strong>Still trading after 6 hours:</strong> 166 coins (21%).</li></ul><p>The median coin had its last trade 14 minutes after launch. In the first hour, the median coin saw 40 transactions on its curve. 288 coins saw fewer than 10.</p><p>15 coins graduated. That sounds like 2%, but 7 of them were not normal launches. In two cases the creator bought 79.3% of the supply in the create transaction itself, which is the whole bonding curve. In the other five, the creator and one or two wallets in the same block filled the curve within a few trades. A sniper could not have bought those coins at a useful price. That leaves 8 coins, 1% of the total, that graduated after real trading by many wallets.</p><p>So the starting point is plain. Without filters, a sniper buys into a pool where about 4 in 10 coins are done within 10 minutes and about 1 in 100 graduates. The job of a filter is to move you toward the better part of that pool.</p><h3>Filter 1: creator history (a blocklist that updates itself)</h3><p>The first thing a bot can check is the wallet that created the coin. It knows this the moment it sees the create transaction, so it costs no time.</p><p>In our two hours, 343 wallets created the 807 coins. 228 of them launched one coin. But 38 wallets launched five or more each, and together they made 378 coins. That is 47% of all launches from 11% of the creators.</p><figure><img alt="Bar chart of Pump.fun coin fate by how many coins the creator launched in two hours: 1 coin, 228 coins, 29% dead within 10 minutes and 41% still trading after 1 hour; 2 to 4 coins, 201 coins, 33% and 33%; 5 or more coins, 378 coins, 53% and 20%. The 3 busiest creators made 99 coins with no creator buy, none still trading after 1 hour." src="https://cdn-images-1.medium.com/max/1024/1*3ZE4VjJnZiTPz8ptnUdN2g.png" /></figure><p>Here is how their coins did, grouped by how many coins the creator launched in the two hours:</p><ul><li><strong>One coin:</strong> 228 coins. 29% dead within 10 minutes. 41% still trading after one hour.</li><li><strong>Two to four coins:</strong> 201 coins. 33% dead within 10 minutes. 33% still trading after one hour.</li><li><strong>Five or more coins:</strong> 378 coins. 53% dead within 10 minutes. 20% still trading after one hour.</li></ul><p>The top three creators are a clear example. One wallet launched 42 coins in 53 minutes. Another launched 32 in about an hour, and a third launched 25. None of their 99 coins included a creator buy, and none of them was still trading after one hour. For most of them, the last trade came in the same minute as the launch.</p><p>There is a catch, and it is why we call this a blocklist and not a hard count. Three of the 8 coins that graduated with real buyers came from busy creators: two from a wallet that launched 17 coins, and one from a wallet that launched 11. A simple rule like “skip any creator with five or more launches today” would have skipped those three too.</p><p>A better rule looks at results, not just the count:</p><ul><li>Skip creators who launched many coins recently and none of them traded past a few minutes.</li><li>Keep a list of wallets you have seen do this, and add to it every day.</li><li>Do not block a creator just because they launch often, if some of their coins did well.</li></ul><p>In DexSnipe this is the blocklist. You can add creator wallets to it by hand, and the bot skips any coin they launch. Whatever bot you use, check that it can block by creator wallet, not only by token name.</p><h3>Filter 2: the creator’s own buy (a minimum and a cap)</h3><p>Most launches on Pump.fun include a buy by the creator in the same transaction. In our data, 526 of 807 creators bought some of their own coin at launch. The median buy was small, about 0.7% of the supply. A bot sees this buy in the create transaction, so this filter also costs no time.</p><figure><img alt="Bar chart of Pump.fun coin fate by creator buy size at launch: no buy, 281 coins, 60% dead within 10 minutes and 17% trading after 1 hour, 2 real graduates; under 2%, 366 coins, 36% and 30%, 5 graduates; 2 to 5%, 58 coins, 28% and 41%, none; 5 to 10%, 54 coins, 17% and 54%, none; 10% or more, 48 coins, 15% and 52%, 1 graduate." src="https://cdn-images-1.medium.com/max/1024/1*nYjMGlIQDcyT798uUgVjqQ.png" /></figure><p>Here is how the coins did by the size of the creator’s buy:</p><ul><li><strong>No creator buy:</strong> 281 coins. 60% dead within 10 minutes. 17% still trading after one hour. 2 graduated with real buyers.</li><li><strong>Under 2% of supply:</strong> 366 coins. 36% dead within 10 minutes. 30% still trading after one hour. 5 graduated with real buyers.</li><li><strong>2% to 5%:</strong> 58 coins. 28% dead within 10 minutes. 41% still trading after one hour. None graduated.</li><li><strong>5% to 10%:</strong> 54 coins. 17% dead within 10 minutes. 54% still trading after one hour. None graduated.</li><li><strong>10% or more:</strong> 48 coins. 15% dead within 10 minutes. 52% still trading after one hour. 1 graduated with real buyers, and 3 were bought out.</li></ul><p>Two things stand out.</p><p><strong>A creator who puts in nothing usually walks away fast.</strong> Coins with no creator buy were the most likely to die in minutes. This group includes most of the coins from the busiest creators, which makes sense. If you launch 42 coins in an hour, you do not buy any of them.</p><p><strong>A large creator buy keeps the coin active, but it did not lead to graduation.</strong> Coins where the creator took 2% to 10% stayed active much longer. But across 112 coins in that range, none graduated. A big creator bag gives the creator a reason to keep the coin moving, and it also gives them a lot to sell. More activity can simply mean the creator, or bots, are trading against new buyers.</p><p>So this filter works best with two limits:</p><ul><li><strong>A minimum.</strong> Skip coins where the creator bought nothing. It removes a big chunk of the fastest deaths.</li><li><strong>A cap.</strong> Skip coins where the creator holds more than you are comfortable with. In DexSnipe this is the dev holding cap. Something between 5% and 10% is a common place to start.</li></ul><p>Seven of the 8 coins that graduated with real buyers had a creator buy of 0.5% of supply or less. The quiet start was more common among real graduates than the big one.</p><h3>Filter 3: buyers in the creator’s block (bundle check)</h3><p>Some launches come with other wallets buying in the same block as the create. Often these wallets are run by the creator, and the buys are sent together as a bundle so they land first. A bundle check counts how many wallets other than the creator bought in that first block.</p><p>In our data:</p><ul><li><strong>No other buyers in the first block:</strong> 676 coins. 45% dead within 10 minutes. 25% still trading after one hour. 6 graduated with real buyers.</li><li><strong>One or two other buyers:</strong> 54 coins. 26% dead within 10 minutes. 46% still trading after one hour. 1 graduated with real buyers, and 5 were bought out.</li><li><strong>Three or more other buyers:</strong> 77 coins. 18% dead within 10 minutes. 53% still trading after one hour. 1 graduated with real buyers.</li></ul><p>The pattern is the same as with large creator buys. Bundled launches look alive. Coins with three or more same-block buyers saw a median of 2,852 trades in the first hour, against 24 for coins with none. But only one of those 77 coins graduated with real buyers. And 5 of the 7 buyouts came from launches with one or two same-block buyers.</p><p>A busy chart in the first hour is easy to read as demand. Our data cannot tell who made those trades, but it does show that busy bundled launches graduated no more often than quiet ones. If your plan is to hold a coin and ride it toward graduation, a bundle check removes coins where a few wallets already sit on a big share before you can buy.</p><p>One problem: a bot can only count same-block buyers after that block lands. We explain what that means for timing further down.</p><h3>Filter 4: top 10 holder share</h3><p>The top 10 holder filter adds up the share of supply held by the 10 biggest wallets, not counting the bonding curve itself. We measured it at slot 25, about ten seconds after launch.</p><ul><li><strong>Under 5%:</strong> 668 coins. 46% dead within 10 minutes. 25% still trading after one hour.</li><li><strong>5% to 15%:</strong> 62 coins. 26% dead within 10 minutes. 45% still trading after one hour.</li><li><strong>15% to 30%:</strong> 61 coins. 13% dead within 10 minutes. 56% still trading after one hour. 2 graduated with real buyers.</li><li><strong>30% or more:</strong> 16 coins. 5 of them were bought out in seconds.</li></ul><p>At this early point most coins have very few holders, so the top 10 share is low almost by default. The filter is useful at the top end. A coin where 10 wallets hold 30% or more after ten seconds is often a coin that a small group has already taken. In our data, 5 of the 7 buyouts sat in that group.</p><p>In DexSnipe the top 10 holder cap is a separate setting from the dev cap. Around 25% to 35% is a reasonable place to start. Set it too low and you skip almost every coin with any early buying.</p><h3>Filter 5: where the metadata is hosted</h3><p>This one is less known. Every Pump.fun coin has a metadata file with its image, description and links. The create transaction points to it with a link. We grouped coins by the host in that link.</p><ul><li><strong>ipfs.io:</strong> 511 coins (63%).</li><li><strong>metadata.j7tracker.io:</strong> 139 coins.</li><li><strong>meta.uxento.io:</strong> 55 coins.</li><li><strong>m.rapidlaunch.io:</strong> 28 coins.</li><li><strong>Other hosts:</strong> 74 coins.</li></ul><p>The non-ipfs.io hosts belong to third-party tools built for launching coins fast, some of them straight from a tweet. Together they made 296 coins, 37% of launches. All 8 coins that graduated with real buyers had their metadata on ipfs.io. None of the 296 coins from the other hosts did.</p><p>The death rates were similar, though. Coins from j7tracker and rapidlaunch links were dead within 10 minutes about as often as the rest. Eight graduates is a small number, so we would not build a strategy on this alone. But a launch tool host is free to check at slot 0, and it is a useful extra signal if you only want to buy coins that someone took the time to set up by hand. If your bot lets you filter on the metadata link, it is worth testing.</p><h3>What did the 8 real graduates have in common?</h3><p>Only 8 coins graduated after real trading. That is too few for statistics, but they are worth looking at one by one, because they are the coins a sniper actually wants.</p><ul><li><strong>Small creator buys.</strong> Five of the 8 had a creator buy under 0.5% of supply. Two had none. One had 17.3%.</li><li><strong>Quiet first block.</strong> Six had no other buyers in the creator’s block.</li><li><strong>Low early concentration.</strong> Six had a top 10 share of 1% or less after ten seconds.</li><li><strong>Normal metadata.</strong> All 8 used ipfs.io links.</li><li><strong>Mixed creators.</strong> Four came from creators who launched only one coin. Three came from creators who launched 11 or 17.</li><li><strong>Real trading.</strong> They saw between 255 and 6,000 trades on the curve in the first hour.</li></ul><p>In plain words, the coins that made it mostly looked boring at launch. Small creator buy, no bundle, no big holders. The loud launches, with big creator buys and crowded first blocks, looked busier in the first hour but almost never graduated.</p><p>That is useful for a sniper, because the boring profile is also the cheapest to buy. Nobody is fighting you for the first block on a quiet launch, so you do not need a large tip to get in. Our post on <a href="https://medium.com/@dexsnipecom/how-to-snipe-meme-coins-on-solana-what-2-023-replayed-trades-say-about-timing-fees-and-exits-524c50d4e00e">how to snipe meme coins on Solana</a> shows what happens to returns when you pay a lot to be first.</p><h3>When can a sniper bot check each filter?</h3><p>This is the part most guides skip. Filters are not free. Some can be checked the moment the coin appears. Others need the coin to trade first, and every slot you wait is about 0.4 seconds of price moving away from you.</p><figure><img alt="List of Pump.fun sniper bot filters and when a bot can check them: creator blocklist, creator buy minimum and cap and metadata link host at slot 0; same-block buyers and top 10 holder cap from slot 1; socials need an outside fetch. 47% of launches came from creators with 5 or more coins, 35% had no creator buy, 37% used a launch tool host, 10% had 3 or more same-block buyers, 2% had 30% or more in the top 10." src="https://cdn-images-1.medium.com/max/1024/1*Y6WblLYbvU3RWf5Z03jNyQ.png" /></figure><p><strong>Checks you can run at slot 0, with no delay:</strong></p><ul><li><strong>Creator wallet and blocklist.</strong> The creator is in the create transaction.</li><li><strong>Creator buy size.</strong> The creator’s buy is in the create transaction too.</li><li><strong>Metadata link host, name and symbol.</strong> These are arguments of the create instruction.</li></ul><p><strong>Checks that need at least one more block:</strong></p><ul><li><strong>Same-block buyers.</strong> You only know who else bought in the first block after it lands.</li><li><strong>Top 10 holder share.</strong> This changes with every trade, so the number at slot 1 is not the number at slot 10.</li></ul><p><strong>Checks that need you to fetch something:</strong></p><ul><li><strong>Socials in the metadata file.</strong> The file sits on an outside host. Fetching it takes time and can fail. When we fetched metadata files for this study, many requests were rate limited.</li></ul><p>What this means in practice: if you want to land in slot 1 or 2, you can only use the slot 0 checks. In our data those checks alone (blocklist, creator buy minimum and cap, metadata host) already cover most of what the filters can do. Creator history and no creator buy were the two strongest signals for fast deaths, and both are free at slot 0.</p><p>If you are willing to buy a few seconds later, you can add the bundle check and the top 10 cap. Our post on <a href="https://medium.com/@dexsnipecom/solana-sniper-bot-who-really-buys-first-we-traced-411-early-pumpfun-buys-6cc602a342c2">who buys first</a> found that most early buyers already land around slots 6 to 25, so waiting a few slots is normal, not a big handicap.</p><h3>Where did early snipers actually buy?</h3><p>In our earlier study we traced 411 buys that filled in the first ten seconds of these same coins. So we can check which coins those snipers chose.</p><ul><li><strong>129 of the 411 fills (31%)</strong> went into coins from creators with five or more launches.</li><li><strong>191 fills (46%)</strong> went into coins where the creator bought 5% or more.</li><li><strong>154 fills (37%)</strong> went into coins with three or more buyers in the creator’s block.</li><li><strong>185 fills (45%)</strong> went into coins with metadata from a launch tool host.</li></ul><p>Early snipers were drawn to the loud launches. That is not surprising. A bot that watches for activity, or that copies other wallets, will find the busy coins first. But the busy coins are exactly the group that, in our data, almost never graduated with real buyers. If your goal is a quick in and out in the first minute, busy launches can work. If your goal is to hold through the early chaos, the data points the other way.</p><h3>Filter settings we would start with</h3><p>These are starting points, not a recipe for profit. Test them small and change them as you learn.</p><ol><li><strong>Blocklist creators who spam launches with no results.</strong> Add wallets that launched many coins in a day with none trading past a few minutes. Review the list weekly.</li><li><strong>Require a creator buy.</strong> Skip coins where the creator bought nothing at launch. In our data this removed 281 coins, of which 60% died within 10 minutes.</li><li><strong>Cap the creator’s share at 5% to 10%.</strong> This removes the large creator bags that kept coins busy but did not graduate.</li><li><strong>Skip launches with three or more same-block buyers,</strong> if you can wait one block.</li><li><strong>Cap the top 10 holders at about 30%,</strong> if you can wait a few slots.</li><li><strong>Keep your buy size small and your exits set.</strong> Even with every filter on, most coins still end within the hour. Your exit rules matter as much as your entry rules.</li></ol><p>When we applied a simple mix (creator under five launches, fewer than three same-block buyers, top 10 under 30%), 363 coins passed. Of those, 33% were dead within 10 minutes, against 48% of the coins that failed. That is a real improvement. It is not magic. Two out of three coins that passed still stopped trading within the hour.</p><p>For the buy settings themselves, slippage, priority fee and tip, see our post on <a href="https://medium.com/@dexsnipecom/best-sniper-bot-settings-for-pumpfun-slippage-priority-fee-and-tip-from-807-launches-76409e7a5040">sniper bot settings for Pump.fun</a>. It uses the same 807 coins.</p><h3>Filters that do not matter on Pump.fun</h3><p>Many sniper bots come with filters built for any Solana token. Some of them do nothing on Pump.fun, because every coin is made by the same program with the same rules.</p><ul><li><strong>Mint authority and freeze authority.</strong> Pump.fun coins are created with a fixed supply, so these checks return the same answer for every coin. They matter on other launchpads and on pools created by hand.</li><li><strong>Transfer tax.</strong> Pump.fun coins do not have a transfer tax. The check matters for tokens created elsewhere.</li><li><strong>Minimum liquidity.</strong> On the bonding curve, the starting price and depth are the same for every coin. Liquidity filters make sense after graduation, or on other DEX pools.</li></ul><p>If you snipe on more than one launchpad, keep these filters on. Our post comparing <a href="https://medium.com/@dexsnipecom/solana-launchpads-compared-the-hidden-rules-of-letsbonk-jupiter-studio-bags-boop-believe-and-87cef7c5a86d">Solana launchpads</a> shows where they make a difference.</p><h3>Common filter mistakes</h3><ul><li><strong>Filtering on activity.</strong> High trade counts in the first minutes were common in coins with large creator buys and bundles. They were not a sign of a coin that would graduate.</li><li><strong>Waiting for socials.</strong> Fetching the metadata file takes time and fails often. By the time it loads, the first buyers are in. Use the slot 0 checks first.</li><li><strong>Setting the top 10 cap too low.</strong> At ten seconds, a top 10 cap of 5% skips almost every coin that has any buyers at all.</li><li><strong>Blocking creators by count only.</strong> Some busy creators did launch coins that graduated. Look at what their coins did, not just how many there were.</li><li><strong>Forgetting the exit.</strong> Filters change which coins you buy. They do not change the fact that most coins end within the hour. Set take profit, stop loss and a time limit before you start.</li></ul><h3>How to choose a Pump.fun sniper bot with good filters</h3><p>If you are comparing bots, these are the questions we would ask, based on what the data showed:</p><ol><li><strong>Can it block creator wallets?</strong> Creator history was the strongest single signal. A bot that only filters by name or ticker misses it.</li><li><strong>Can it set a minimum and a maximum creator buy?</strong> Both ends mattered in our data.</li><li><strong>Does it have a bundle check and a top 10 cap,</strong> and can you turn them off when you want the fastest entry?</li><li><strong>Does it run the filters before the buy, on a server?</strong> A bot that runs in your browser or waits for you to click cannot buy in the first seconds.</li><li><strong>Are the exits part of the same setup?</strong> Take profit, stop loss and a timer should be set once and run without you.</li><li><strong>What does it charge per trade?</strong> Good filters cut the number of trades. A flat price or a low fee rewards that. A percentage fee on every buy and sell adds up fast if your filters are loose.</li></ol><h3>Questions people ask about Pump.fun sniper bot filters</h3><h4>What is the best filter for a Pump.fun sniper bot?</h4><p>In our data, creator history. Coins from wallets that launched five or more coins in two hours were dead within 10 minutes 53% of the time, against 29% for one-time creators. Requiring a creator buy was a close second.</p><h4>Do filters help you find coins that graduate?</h4><p>Not much. Only 8 of 807 coins graduated with real buyers, and no filter picked them out cleanly. Filters are good at removing coins that die fast, which saves money on fees and bad entries.</p><h4>Should I skip coins where the creator bought a lot?</h4><p>Coins with a creator buy above 5% stayed active longer, but only one of 102 graduated with real buyers. A dev holding cap of 5% to 10% is a sensible default.</p><h4>What is a bundled launch on Pump.fun?</h4><p>A launch where other wallets buy in the same block as the create transaction. Usually these buys are sent together with the create. In our data, 77 coins had three or more such buyers.</p><h4>Can a sniper bot check socials before buying?</h4><p>It can, but it is slow. The links sit in a metadata file on an outside host, and fetching it takes time and can fail. Checks from the create transaction itself are instant.</p><h4>How many new Pump.fun coins graduate?</h4><p>In our two hours, 15 of 807 graduated within a day. Seven of those were bought out by the creator or one or two wallets in seconds. Eight, about 1%, graduated after real trading.</p><h4>Do mint and freeze authority filters matter on Pump.fun?</h4><p>No. Every Pump.fun coin is created the same way, so these checks return the same result for all of them. They matter on other launchpads.</p><h4>Is a top 10 holder filter useful in the first seconds?</h4><p>Only at the high end. Most new coins have very few holders at first, so the share is low by default. A cap around 30% catches coins that a small group has already taken.</p><h3>Sources and further reading</h3><ul><li>Pump.fun public program docs and IDL: <a href="https://github.com/pump-fun/pump-public-docs">github.com/pump-fun/pump-public-docs</a></li><li>Pump.fun docs, bonding curve trades v3: <a href="https://github.com/pump-fun/pump-public-docs/blob/main/docs/instructions/TRADE_V3.md">github.com/pump-fun/pump-public-docs</a></li><li>Solana docs, terminology (slots and blocks): <a href="https://solana.com/docs/references/terminology">solana.com/docs/references/terminology</a></li><li>Helius docs, getTransactionsForAddress: <a href="https://www.helius.dev/docs/rpc/gettransactionsforaddress">helius.dev/docs</a></li><li>Jito docs, bundles and low latency transaction send: <a href="https://docs.jito.wtf/lowlatencytxnsend/">docs.jito.wtf/lowlatencytxnsend</a></li><li>Metaplex docs, token metadata standard: <a href="https://developers.metaplex.com/token-metadata/token-standard">developers.metaplex.com</a></li><li>IPFS docs, what an IPFS gateway is: <a href="https://docs.ipfs.tech/concepts/ipfs-gateway/">docs.ipfs.tech</a></li><li>Solscan block explorer, to check any coin or creator yourself: <a href="https://solscan.io/">solscan.io</a></li></ul><p>All numbers come from our own pull of Solana mainnet data for 807 Pump.fun coins created between 02:05 and 04:05 UTC on November 10, 2026, followed for 24 hours. Nothing here is financial advice. Meme coins are very risky, and most new coins go to zero.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5c7a932f9e9e" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>