January 07, 2026

Updated computer vision model and geomodel with over 1,000 new taxa

Happy New Year evreryone! We released a computer vision and geomodel update v.2.27 today with 112,613 taxa compared to 111,435 taxa in the last model and is trained off data exported on November 16, 2025.

This model is the result of millions of observations and identifications shared on iNaturalist. As new taxa are observed and identified, taxa (mostly species, but also genera, families, etc.) are added to the model once there are around 100 photos and 60 observations of that taxon.

As taxonomy changes and misidentifications get corrected by the community, sometimes taxa are also removed from the model. To keep up with these changes, we update the model every month or two so that the community can benefit from the improvements. Check out our help page for the most updated information on how we update the model.

Below are links to the taxa added since the last model. You can click through and search for your username to see if you have observed or identified any of these species. If so, thank you!

The graph below shows how the number of taxa included in the model has grown over the last few years — more than doubling from 55,000 in 2022 to over 110,000 today.

Each time we release a new model, we evaluate it against the previous one.

The graph below shows model accuracy estimates using 1,000 random Research Grade observations in each group not seen during training time. The paired bars below compare average accuracy of model 2.26 with the new model 2.27.


Thank you to everyone in the community who contributed observations and identifications for all the species in this model, and to everyone continuing to share feedback with us as we improve it further! This collective effort wouldn't be possible without you.


Posted on January 07, 2026 07:23 PM by loarie loarie | 8 comments | Leave a comment

December 17, 2025

A new orchid species, found through the power of community science

A chance encounter led to a scientific breakthrough for biologist David Alejandro Sánchez Gómez (@davidsanchez10), Yudy Gallego (@yudyalejag), and collaborators. "In 2019, during a trip to a páramo in the municipality of Sonsón (known locally as Las Palomas), I photographed an orchid and uploaded it to iNaturalist," says David, who calls El Carmen de Viboral, Colombia home. "That observation turned out to be very special."

The iNaturalist community was pivotal when it came to setting the record straight; this was a never-before-described species. "Through the platform, Mark Wilson contacted me and suggested we collaborate to describe it as a new species," says David. "Since I wasn't an orchid expert, I invited my colleague Yudy Gallego, a national specialist in orchids, and together we carried out multiple field visits, photographed the plants, studied their habitat, and documented their flowering. That teamwork eventually led us to formally describe the species."

Very special and very vulnerable

Pleurothallis maitamae, the newly described orchid, is as distinctive and unique as the high altitude ecosystem it calls home. Keeping tabs on its populations could help shed light on the health of its critical habitat.

"In Colombia, páramos are unique high-mountain ecosystems located above the continuous forest line and below the glaciers," says David. "They are globally important because they provide fresh water to millions of people, host an extraordinary diversity of plants and animals, and act as natural regulators of the climate. Unfortunately, they are also highly threatened by agriculture, cattle ranching, mining, and climate change."

"Within this fragile context, what I find most fascinating about Pleurothallis maitamae — also known locally as the Maitama Orchid — is that it is restricted to a single páramo fragment in Sonsón," continues David. "That makes it both very special and very vulnerable. Despite previous botanical studies in the area, it had never been noticed before. It's also a very striking orchid that blooms only once a year, between July and August, which makes it even harder to find and even more remarkable to study."

As Yudy shares, "the genus Pleurothallis is particularly fascinating to me because of its huge diversity and morphological variation. In the case of the species we described, what captivated me most was its labellum, a finely ornamented structure so intricate and delicate that it can only be truly appreciated under a magnifying lens or stereoscope. It's amazing how a plant can develop such tiny, complex forms in service of its pollination."

To capture the orchid in full detail, researchers had to be resourceful. "One of the main challenges we faced was photographing such small floral structures without specialized equipment," says Yudy. "We had to collect a few flowers, keep them in a damp cloth, and get up early the next morning to work in the lab — to dissect the flowers and document them before they dried or became damaged. It was a very careful but deeply rewarding process."

"When we confirmed it was an undescribed species, we felt an incredible mix of joy and responsibility," says David. "It was our first time going through the process of describing a species, and that can be intimidating. At the same time, it was very motivating because we knew the description was also an important step toward future conservation actions, given the threats that the species and its habitat face."

"Anyone could have uploaded those photos"

"What is really striking to me is that when we first came across this orchid, we were just early undergraduate students," says David. "In other words, anyone could have uploaded those photos — and in fact, several other iNaturalist users (at least seven people) had also documented the same orchid. It was the collective perspective of these records that gave us the confidence to realize we were looking at something new."

Photos that may seem casual can hold great conservation value. As Yudy shares, "it was truly striking to realize that a plant simply photographed in the field turned out to be a unique record for the region and for the country."

So continue taking those photographs and adding them to iNaturalist! And a tip for the perfectionists out there: "Don't be afraid to upload your photos, even if they're not 'perfect,'" says David. "Follow your curiosity about the natural world — biodiversity is everywhere, not just in pristine forests but also in gardens, farms, and urban areas. With just a simple photo, you can contribute to conservation and connect with a global community that shares your passion."

The power of community science

Community science is empowering — "It allows people to be active participants, not just recipients of information, but true partners in the care and appreciation of the natural wealth of their territories," says David.

"For me," says David, "community science means the democratization of scientific knowledge. It's about building global networks of naturalists, fostering collaboration, and creating opportunities for anyone to contribute to biodiversity research. I also see it as a way of learning through photography, not only from a scientific perspective but also by appreciating the beauty of the species around us."

"In the future, I hope to develop projects where local communities not only collect data but also ask questions, generate hypotheses, analyze results, and share outcomes," says David, who is part of Viboral Biodiverso, an organization which shines a light on the rich biodiversity in eastern Antioquia, in Colombia. "That way, community science can expand from simple observation to becoming a more complete and empowering research process."

"I'd also like to highlight that this is not the only case in Colombia where iNaturalist has helped connect people and contributed to the discovery of new species. I think it's really important to give more visibility to these stories, as they show how the platform is actively shaping biodiversity research and conservation.

"One more piece of advice I would share, especially with biologists, is to spend more time exploring iNaturalist," says David. "Review observations, and if you are an expert in a particular group, take the time to help others learn more about the species in their territory. Who knows — in the process, you might even stumble upon a new discovery yourself."

Posted on December 17, 2025 08:20 PM by seastarya seastarya | 8 comments | Leave a comment

Provisional Taxa Pilot

We’re launching a new pilot that allows Provisional Taxa to be added to iNaturalist under specific constraints. This pilot is starting small and focused, but it marks an important step toward better supporting DNA-based species discovery, especially in fungi.

What’s included in the pilot

For now, provisional taxa are allowed only within the family Cortinariaceae and can only be made by staff, with short-term plans to allow curators to create provisional taxa and to expand to additional fungal groups.

Norms for creating a provisional taxon:

  • It must be supported by an observation that includes the “DNA Barcode ITS” Observation Field.
  • A link to the earliest supporting observation supported by the corresponding “DNA Barcode ITS” Observation Field should be added in the “Why are you making this change?” field when creating the taxon, which appears in the taxon’s history.
  • Provisional names must follow the format:
    Genus sp. 'Label'
    Example: Calonarius sp. 'CA02'
    • The genus-like part may contain letters only.
    • The label in quotes may contain letters, numbers, and hyphens.

How provisional taxa behave


Provisional taxa function similarly to regular taxa, but with a few key differences:

  • On web taxon pages, they are marked with a “[Provisional]” badge.
  • In the taxon page’s tree, provisional taxa are hidden by default but can be revealed by toggling “Show provisional taxa.”
  • Observations of provisional taxa are not shared with external biodiversity archives such as GBIF.

What was changed today

Today, we deployed new functionality supporting provisional names (with the creation of provisional names currently restricted to staff) and created a Provisional Taxon for Calonarius sp. 'CA03'. We also automated the creation of 10 identifications matching the Provisional Names Observation Field. For example, from the Provisional Name Observation field by @ssquazzo with the value "Calonarius sp. 'CA03'" found on this observation, we automated the creation of a corresponding identification by ssquazzo of the new Provisional Taxon Calonarius sp. 'CA03'.


Be aware that anyone can now add IDs of Calonarius sp. 'CA03' as if it were a regular taxon. Our plan is role out the creation of more Provisional Taxa within family Cortinariaceae and corresponding Identifications based on these Provisional Name Observation Fields in the coming days.

Why we’re making this change

The fungi community on iNaturalist has been at the forefront of using DNA sequencing to understand mushroom diversity. Because fungal taxonomy is still developing—and many widely observed mushrooms turn out to be undescribed species — users have been tracking these unnamed lineages using Observation Fields. While effective, this approach is cumbersome and limits discoverability. We’ve worked on this solution in close collaboration with @ym_wang_pnw, @dannymi, @alan_rockefeller, @malacothrix, and @stevilkinevil.

By supporting Provisional Taxa, we hope to make it easier to organize, discuss, and learn from DNA-based lineages while maintaining scientific rigor and clarity on the platform.

This pilot is an early step toward understanding how iNaturalist can help document the many species that remain undescribed, and how DNA sequencing and community science can work together to illuminate the vast unknowns of fungal diversity.

Posted on December 17, 2025 01:37 AM by loarie loarie | 73 comments | Leave a comment

December 15, 2025

ID-a-thon starts today!

December 15, 2025 – January 15, 2026

It’s officially time to kick off the first iNaturalist ID-a-thon: a month-long celebration of identifying observations and helping each other learn more about the living world. Whether you’ve made zero IDs, a few hundred IDs, or tens of thousands of IDs, this is your moment to jump in, sharpen your skills, and make a real impact for the global iNaturalist community.


What We’re Aiming For

The ID-a-thon has three goals:

  • Increase the number of active identifiers, especially newer ones
  • Increase overall identification activity during this time, giving more community members accurate IDs on their observations
  • Grow our collective expertise by offering fun, accessible weekly IDing activities

Every accurate identification helps someone learn, helps an observation progress, and helps researchers and conservation organizations access higher-quality data. IDs are a critical part of iNaturalist, and everyone has ways they can contribute!

Learn more about the ID-a-thon & making identifications


What to Expect This Month

From now through January 15, we’ll post a new weekly IDing activity designed to help you build your skills and discover new groups of organisms to identify — no expert knowledge required. Our first activity starts today: keep reading to find out more!

🏅 Best Practices for Identifying

  • Make sure you can support your ID based on evidence and information you can cite if asked.
  • Only identify to the taxonomic level the evidence supports — blurry or distant images might only warrant "Plants" or "Birds" or "Insects."
  • Check your notifications to see if anyone agrees or disagrees with your identifications. If it looks like you might have misidentified something, you can withdraw your identification.
  • It's always okay to skip! If you're unsure, move on to the next observation.
  • Identifying is easiest on the iNaturalist website, where you can use the Identify Page and see larger images.

Week 1 Activity: Identifying “Unknowns”

Some observations come in with no ID at all, which lists them as “Unknown.” Helping sort these into even broad categories (plants, insects, fungi, etc.) immediately makes them more likely to be seen and identified by others. It’s one of the easiest and most impactful ways to contribute during the ID-a-thon. If you’re new to identifying, this is a perfect place to start.


Learn how to add identifications to unknown observations



📈 Tracking the Impact: New Daily Stats

To help everyone follow along with the ID-a-thon’s progress, we’ve added new identifier and identification metrics to the iNaturalist Stats page. If you want to watch the effect of your work, keep an eye on:

Identifiers on the Daily Users graph (the yellow line): this shows the number of people each day who have added identifications to other people’s observations — even just one identification a day will count you on this graph!

Identifications for Others (1 day) on the Identifications graph (the green line): this shows the total number of identifications that have been made on other people’s observations each day — so if two people add an ID to an observation (e.g. one person IDs an unknown observation as “Plants” and then a second person comes in and IDs it down to species), that adds two identification counts to this graph.

The Observations Identified By Others (1 day) graph (the left graph, with just the blue line): this shows the total number of observations that have had IDs added to them each day; in the previous example of two people adding an ID to an observation, that adds one identified observation count to this graph.

These numbers will shift noticeably if the community jumps in together — so let’s move them! These graphs update once per day, so be sure to check in daily.


Join the ID-a-thon!

We’d love for you to be part of this month-long push to help each other learn and improve the quality of biodiversity data around the world.

🎯 Tell us your goal!

In the comments below, share the challenge or commitment you’re setting for yourself during the ID-a-thon:

  • Adding at least 1 identification each day?
  • A certain number of IDs per day?
  • Trying to learn how to ID a new taxon?
  • Clearing unknowns from a specific area?
  • Helping new observers?

Whatever it is, we can cheer each other on.

Let’s make this month a huge boost for the iNaturalist community, and have fun while we’re at it. Happy identifying!

Posted on December 15, 2025 03:09 AM by kestrel kestrel | 120 comments | Leave a comment

December 11, 2025

Revisiting “Based on the evidence, can the Community Taxon be improved?”

Background

We recently found a bug that was resulting in some cases where the Research Grade observations did not have community support. The definition of Research Grade has always required community support for the observation taxon (the node in the taxonomy where observation sits), so fixing this bug was a priority.

Here’s an example of the bug. This sequence resulted in a Research Grade observation at the subspecies rank even though there was no community support beyond species:

After fixing the bug by requiring that in all Research Grade observations the Community Taxon and Observation Taxon match, this observation would now be Needs ID:

We know that many identifiers would like observations to become Research Grade. Here’s what you can do to help make this observation Research Grade. First, investigate whether the observation can be identified to subspecies. One way is by asking Identifier B to share their reasoning:

If you disagree with their reasoning, you can add a disagreeing ID at the species level which will roll it back to Research Grade at the species level:

If you’re convinced by their reasoning, you can add a supporting subspecies ID which will make the observation Research Grade at the subspecies level:

Edge case: ID order is important for subspecies

Be aware that iNaturalist handles infraspecies nodes like subspecies a bit differently than other nodes. If the order of the first two IDs in the above example were switched (i.e. if the first identifier added species ID and the second identifier added a subspecies ID), the observation won’t roll forward to subspecies until a third identifier adds a subspecies ID. We did this because many observers don’t want their (often Research Grade) species-level observations to become Needs ID observations at the subspecies-level after just a single subspecies identification.

Changes to “Based on the evidence, can the Community Taxon be improved?”

In general observations become Research Grade when they have community support at the species level. The “Based on the evidence, can the Community Taxon be improved?” Data Quality Assessment (DQA) metric was designed to allow people to kick observations out of Needs ID if they can never reach the species level (photo blurry or missing diagnostic characters etc.). If the observation would otherwise be Research Grade and the Community Taxon is finer than family-level, the observation gets kicked into Research Grade. Otherwise, it gets kicked into Casual.

The “Based on the evidence, can the Community Taxon be improved??” DQA metric should be used sparingly. There are currently about 20,000 observations kicked into casual because of misuse of this metric. We’ll describe them here and then explain a few minor changes we made to this metric to try to constrain its use. You can load them into the identify tool with newly added URL search filters for failing DQA metrics .

Common misuses of the metric:

1. Forgetting that DQA the metric refers Community Taxon, not the Observation Taxon

Sometimes identifiers add an ID that they don’t think can be improved and then vote no to “Based on the evidence, can the Community Taxon be improved?”. Remember that the metric refers to the Community Taxon which may still be sitting at a coarser node. Since the Community Taxon and Observation Taxon don’t match, this observation will get kicked into Casual.

If you see observations in this state, please vote “yes” to “Based on the evidence, can the Community Taxon be improved?” so they can get more scrutiny by people who can hopefully roll the Community Taxon forward.

2. Voting “no” rather than adding a disagreeing ancestor ID

Sometimes the community adds a coarser non-disagreeing ID and votes “no” to “Based on the evidence, can the Community Taxon be improved?”. Because the Community Taxon and Observation don’t match, this kicks the observation into Casual.

Rather, add a disagreeing ID to roll the Community Taxon back so that it matches the Observation Taxon. Then it will still be eligible for Research Grade if you want to kick it out of Needs ID with a “no” vote.

3. Adding a finer ID to an observation kicked out of Needs ID without a “yes” vote

Sometimes identifiers add finer IDs to observations that have been intentionally kicked out of Needs ID with a “no” vote (often into Research Grade). Because the Observation Taxon no longer matches the Community Taxon, this kicks the observation into Casual.

Rather, accompany your finer ID with a “yes” vote to bring the observation back into Needs ID.

You should also in these cases explain why you disagree with the previous identifier and think that the Community Taxon can be improved.

Changes to “Based on the evidence, can the Community Taxon be improved?” Metric

We made the following changes to the metric to try to constrain its use to minimize confusion and misuse.

1. Changed the name of the metric

We changed the name of this metric from “Based on the evidence, can the Community Taxon still be confirmed or improved?” to “Based on the evidence, can the Community Taxon be improved?” because the intention of this DQA metric is that the Community Taxon cannot be improved (e.g. improved to a finer node) but not necessarily that it can’t be confirmed (e.g. an agreement at the current node).

2. Resetting votes when the Community Taxon changes

We’ve realized that the votes added to this metric often become out of context as the Community Taxon they referred to later changed. This makes interpreting this metric very confusing. As a result, we made a change so that the “Based on the evidence, can the Community Taxon be improved?” votes reset when the Community Taxon changes. For example, as shown in the sequence below, Identifier B’s “No” vote referring to refining the genus, is reset once the Community Taxon rolls back to Tribe to make it clear that the reference was to refining from Genus but not necessarily from refining from Tribe.

3. Disabling voting when the Community Taxon does not match the Observation Taxon

“Based on the evidence, can the Community Taxon be improved?” is disabled if the Community Taxon does not exist. We also made a change disabling it if the Observation Taxon does not match the Community Taxon. We did this because the Observation Taxon and Community Taxon don’t match only in situations where the identifications haven’t reached a consensus, and we’d prefer that the identifications reach a consensus before resorting to kicking the observation out of Needs ID.

Edge case: Opting out of the Community ID

One last caveat. If the observer chooses to opt-out of the Community Taxon and is unresponsive, the community needs to be able to remove observations from Needs ID that can’t be improved regardless of whether the Community Taxon and Observation Taxon match.

As a result, the “Based on the evidence, can the Community Taxon be improved?” DQA metric is never disabled when the observer has opted out of the Community Taxon.

As a courtesy, when voting one of these observations out of Needs ID, you should add a comment or ID remark mentioning the observer to increase the chances that they will see the status of their observation and hopefully make the required changes.

Posted on December 11, 2025 09:58 PM by loarie loarie | 80 comments | Leave a comment

December 05, 2025

Observation Accuracy Experiment v0.5

Update added at the bottom of this post on Dec 12, 2025
Last year we ran four experiments to better understand the accuracy and quality of iNaturalist data and how to improve it. Today we’ve launched our fifth Observation Accuracy Experiment.

Here’s a reminder about how these experiments work:

Step 1: Sample Generation and Assignment

From the iNaturalist database (A), we generate a random sample of 10,000 observations (B). Subsamples are assigned to candidate validators (C) based on the criteria described below. We aim to assign each sample observation to multiple candidate validators ("Target validator redundancy").

Validator candidate criteria

If an identifier has made at least 3 improving identifications on a taxon in a country, we consider them qualified to validate that taxon within that country. Improving identifications are the first suggestion of a taxon that the community subsequently agrees with.
For example, Identifier 1 adds an ID of A to an observation. If Identifier 2 later adds a leading ID of B, Identifier 1's ID on A becomes an improving ID. If Identifier 3 later adds a supporting ID to B, Identifier 2's ID on B becomes an improving ID.

Step 2: Validating samples

We send emails to each validator with a link to their subsample loaded in the iNaturalist identify tool, instructions to identify each as best they can, and a deadline after which we will use the new identifications to assess the accuracy of the sample. The instructions are for validators to add the finest identification they can to each observation. We know this means that some observations that are already Research Grade might get a flurry of redundant confirming identifications.

Step 3: Scoring sample observations from validations

We score observations in the sample as Correct, Uncertain, or Incorrect by comparing the validator identifications to each sample observation’s taxon. Identifications matching or finer than the sample’s observation taxon are scored as Correct. Non-disagreeing coarser identifications are scored as Uncertain as are scenarios where multiple validators disagree. Disagreeing coarser identifications or identifications on different branches are scored as Incorrect.

Step 4: Accuracy and Precision calculation

Accuracy

The top level statistic we are aiming to estimate is Accuracy (the percent of the sample that is correctly identified). We calculate Accuracy as the percent correct for the entire sample as well as subsets such as the Research Grade subset. We also calculate the percent Uncertain and the percent Incorrect. In the example figure below, the percentage correct would be 83, the percentage uncertain would be 6 and the percentage incorrect would be 11.

Precision

If every sample observation was at a coarse taxon such as kingdom, the accuracy of the iNaturalist dataset might be very high, but the precision might be too low to make the dataset very useful. To estimate precision, we count the number of leaf descendants in the iNaturalist taxonomy for each taxon associated with each sample observation. We calculate precision as 1 divided by the number of leaf descendants. If the taxon is a leaf taxon the precision is 100%. We ignore ranks below species in the precision calculation. In the example figure below the average precision of three sample observations would be 51%.

Thank you so much in advance if we contact you as a potential validator and you choose to participate. We couldn’t do this experiment without your help and expertise!

Update: Dec 12, 2025

Thanks so much to everyone who took part in Observation Accuracy Experiment v0.5 and shared thoughtful feedback. We know this round raised questions about how observations and validators were selected, how scoring works, and what we hope to learn. Below is a quick update on where things landed.

For the Research Grade subset in this experiment, we found:

  • 1% Incorrect
  • 8% Uncertain
  • 91% Correct

“Uncertain” ≠ “wrong.”
“Uncertain” reflects observations we were unable to have reviewed by qualified validators within the experiment’s time window—not that the existing ID was doubtful. Many of the ~450 RG observations in this category are almost certainly correct; they simply weren’t matched to a validator under the criteria we used.

In contrast, 76 RG observations (about 1%) were determined to be incorrect or too finely identified based on the evidence. The remaining ~5,000 (91%) were determined to be correct.

The ~1% incorrect rate aligns well with past experiments. The higher uncertainty compared to previous rounds (often ~3%) may simply reflect that we reached out to about 1,000 fewer validators this time, even though the response rate (60%) was similar.

Things to improve
We’ve also been reading your comments carefully, and your feedback is shaping what we improve next:

  • Validator matching (better taxon + geography sensitivity; distinguishing generalists from specialists)
  • Edge case handling (own observations, missing/obscured locations, unknowns). Many suggested focusing future experiments solely on RG observations.
  • Clearer instructions (use of guides, how to handle well-reviewed or extremely tough observations, and how validator IDs are used)
  • UX issues (notifications, mobile/app behavior, assignment sizes)

We’ll share a fuller write-up of v0.5 once we’ve fully digested the results and feedback. For now, we want to emphasize that:

  • the headline accuracy numbers remain strong, and
  • your participation and critiques are directly informing v0.6 and beyond.

Thank you again for taking the time to participate, reflect, and help us strengthen iNat’s data—and our understanding of it.

Update: Dec 15, 2025

In response to @zoology123's comment below, we completely agree that accuracy varies by taxonomic group and by geography. Its difficult to get accurate estimates for these small groups without much larger sample sizes. But we've updated the all experiment analysis we've shared in the past to include this experiment v0.5. Since we now have over 18,600 Research Grade observations that have been assessed through this experiment process we can start to get reasonably accurate estimates for some pretty niche groups - e.g. based on a sample of 74, our estimate of rare North American Research Grade Insect obs is now 76% (64% - 85%). Keep in mind this analysis collapses across time. This might be evidence that we should focus on Research Grade data to try to more quickly drive larger sample sizes for the groups shown here:

Posted on December 05, 2025 01:34 AM by kestrel kestrel | 91 comments | Leave a comment

December 03, 2025

Scaling identification expertise: Exploring ways to learn from the iNaturalist community

In order to best protect nature, we first need to grow our collective understanding of it, and one way iNaturalist can help is to support more people to learn nature identification skills. Right now, only a small fraction of iNaturalist users actively verify and refine identifications for others. We need more identifiers!

For years, experienced identifiers have been generously sharing their knowledge on iNaturalist — most frequently through making ID remarks and comments on people’s observations. There is a huge amount of knowledge embedded in those comments, but it’s often hard to find. We want to explore using new technology to make this expertise easier to access.

You can now browse the first release of the ID Summaries demo to see species identification tips for about 150 taxa. These tips are drawn from the expertise of about 30 prolific identifiers on iNaturalist.

  1. Every taxon featured has 3–7 ID summaries.
  2. Every ID summary has a photo tip for making new observations of that taxon.
  3. Each ID summary references a set of the original comments used for the summary, so you can always see sources.
  4. You can give feedback on each summary and the associated comments.
  5. Summaries are available in English, Spanish, French, and German.

This demo is separate from the rest of iNaturalist. None of the summaries or feedback in the demo appear anywhere else on the site or in our apps.


Explore the Demo


Your feedback on the demo helps us understand what’s working and what could be improved as we continue exploring this approach. We’re collecting feedback on individual summaries through built-in mechanisms on the demo, and we’d also love to hear your thoughts in the iNaturalist Forum. If you’d prefer to share feedback privately, please feel free to reach out to [email protected].

Watch the webinar

We hosted a webinar (recording below) on this topic, too. Many thanks to everyone who joined us live!

Here’s what you’ll hear about in the recording:

  1. The big picture: Addressing the biodiversity crisis needs more people with the knowledge to identify and understand species.
  2. Preview of the ID Summaries demo: Our first exploration into highlighting, organizing, and summarizing expert identification remarks and comments and making them more findable and accessible.
  3. Community perspectives: Hear directly from @catchang and @nathantaylor, two of the experienced identifiers whose expertise helped shape this demo.
  4. What’s next: Some ideas for future directions of this work, ways you can share feedback, and upcoming identification-related efforts.



Watch the Webinar


Stay tuned:

We’re going to host iNat’s first-ever ID-a-thon very soon — more details to come!

Additional resources

  • How Identifications Work
  • Identifying on iNaturalist webinar
  • When to Agree with an Identification
  • High-Level Identification FAQs

  • Posted on December 03, 2025 07:50 PM by loarie loarie

    December 02, 2025

    New Limpet for iNaturalist! - Observation of the Week, 12/2/25

    Our Observation of the Week is the first Potamacmaea fluviatilis limpet posted to iNaturalist! Seen in Myanmar by @naturalisttuna!

    On October 29th, Jacob Tuna (naturalisttuna) came across some limpet snails as he walked along the shore.

    I was walking in the early morning and spotted those guys on a wood log of an old quay by the river (Pathein River)...that place is intertidal and a mangrove area where brackish waters occur. Even though I am not familiar with these species or family, I clearly knew that they are a type of marine gastropods at the very first glance. So I pulled my phone out of my pocket and took a pic of them. I even tried to extract them off the log to photograph their underside too. Thank God that they were hard as glued or if not, I could have bothered them. Then I did some browsing on the internet and realised that they are indeed marine gastropods called limpets (I've never seen or heard of them before). 

    A few days later iNat’s top limpet identifier, @rlucine, came across Jacob’s observation and was able to identify it as the first one of Potamacmaea fluviatilis posted to the site (and first ever record in Myanmar that will be added to GBIF)! 

    Limpets (subclass Patellogastropoda) are common aquatic gastropods found on rocky shores all over the world, and most prefer full marine habitats. But as Lucie tells me, this species is one of only a few known to prefer estuarine (brackish) areas (like Patelloida mimula). And until now was known only from museum specimens like this one. How does having photos like Jacob’s help us understand this species better? Lucie tells me,

    Having a photo of the living specimen really complements the original description by Blanford here: https://www.biodiversitylibrary.org/page/41274544#page/294/mode/1up 

    He writes: “this species is found on rocks, rarely on trunks of trees, in many of the creeks near high water mark, in brackish water.”

    So having a living specimen with an accurate location pin is basically a “picture is worth 1000 words” scenario, where we can see the habitat demonstrated, contrast the colors and texture of the organism with its environment, and demonstrate exactly how it looks to find an individual. That's great for iNaturalist since we may only see what we know we can find :)

    In this case Blanford also doesn't say whether the organisms are found in a cohort, so the cluster of them that naturalisttuna found is informative. Whether or not limpets form a cohort is sometimes a useful identifying feature, for example Lottia gigantea are solitary, whereas Lottia digitalis cluster, so when presented with ambiguous juveniles or obscured shells we now know we can use clustering as circumstantial evidence to support an identification of this species.

    Jacob’s interest in nature “started with insects and their ecosystem connections. Now, not only insects, I've studied much more broadly. I especially love to study species native to my country (Myanmar) and Southeast Asia regions.” When I asked him how iNaturalist has changed the way he interacts with or sees the natural world, he told me

    Using iNaturalist has changed my perspective on nature. I would rather say “developed my perspectives.” I love this app. Sometimes, searching for insects or whatever in nature and posting them on this app becomes my favourite pastime. And even though it is non-profit, I believe that it has some good impacts on biodiversity surveys or something helpful to biologists or scientists.

    (Some quotes have been lightly edited for clarity.)


    - Jacob has since posted three more observations of Potamacmaea fluviatilis!

    - you can read more about rlucine in this Identifier Profile! She also includes tips on how to photograph them for identification.
    Posted on December 02, 2025 10:32 PM by tiwane tiwane | 15 comments | Leave a comment

    November 24, 2025

    Updated computer vision model and geomodel with over 1,500 new taxa

    We released a computer vision and geomodel update today! Model v.2.26 has 111,435 taxa compared to 109,680 taxa in the last model and is trained off data exported on October 19, 2025.

    This model is the result of millions of observations and identifications shared on iNaturalist. As new taxa are observed and identified, taxa (mostly species, but also genera, families, etc.) are added to the model once there are around 100 photos and 60 observations of that taxon.

    As taxonomy changes and misidentifications get corrected by the community, sometimes taxa are also removed from the model. To keep up with these changes, we update the model every month or two so that the community can benefit from the improvements. Check out our help page for the most updated information on how we update the model.

    Below are links to the taxa added since the last model. You can click through and search for your username to see if you have observed or identified any of these species. If so, thank you!

    The graph below shows how the number of taxa included in the model has grown over the last few years — more than doubling from 55,000 in 2022 to over 110,000 today.

    Each time we release a new model, we evaluate it against the previous one.

    The graph below shows model accuracy estimates using 1,000 random Research Grade observations in each group not seen during training time. The paired bars below compare average accuracy of model 2.25 with the new model 2.26.


    Thank you to everyone in the community who contributed observations and identifications for all the species in this model, and to everyone continuing to share feedback with us as we improve it further! This collective effort wouldn't be possible without you.


    Posted on November 24, 2025 10:48 PM by loarie loarie | 7 comments | Leave a comment

    November 18, 2025

    The iNaturalist team continues growing!

    We're excited to welcome seven new members to the iNaturalist team this year — five on staff, and two on our board of directors.


    New Staff Members

    Abbey Campbell

    Abbey Campbell, Software Engineer


    Abbey joined iNaturalist as a software engineer in October to maintain and expand on our mobile applications. She's been an iNaturalist community member since 2022, with a special interest in California native plants and all the other living things that rely on them. Abbey has a background in fine art as well as extensive experience with building mobile apps on small, agile teams. When she's not creating software, she's making other things: weaving on a floor loom, cooking elaborate dinners, and tackling home improvement projects.


    Arya Natarajan

    Arya Natarajan, Senior Communications Manager


    Arya joined iNaturalist as our first Senior Communications Manager in April, and she's been a community member since 2017. She brings together her background in biology and art as a science communicator, and one of her biggest focuses is to highlight the incredible science and conservation efforts that emerge from the iNat community. Previously, she worked on digital strategy and content at the California Academy of Sciences, where she supported community science campaigns like the City Nature Challenge and Snapshot Cal Coast.


    Seth Peterson

    Seth Peterson, Software Engineer


    Seth joined iNaturalist as a software engineer in October. He brings his background and deep experience as a React Native developer to work on maintaining and expanding on our mobile applications. When he's not at his desk, you'll find him building furniture or coordinating events at Twin Cities Maker, a local makerspace and community workshop.


    Dan Rademacher

    Dan Rademacher, Head of Product and Engineering


    In December, Dan Rademacher will transition from the iNaturalist Board to become our first Head of Product and Engineering. For more than a decade, Dan has managed designers, developers, and data analysts creating web applications, data visualizations, and analytical products for park planning, climate adaptation, equitable public funding, redistricting, and more. He also edited a regional nature magazine for 10 years and has been part of the iNat community since 2009.


    As co-organizer of the Nerds for Nature meetup, in 2013 Dan helped develop the model for grassroots iNaturalist-powered bioblitzes that inspired the City Nature Challenge. Dan most recently served as executive director of GreenInfo Network and previously worked at Stamen Design and Bay Nature Institute.


    Ryan Stelly

    Ryan Stelly, Software Engineer


    Ryan joined iNaturalist as a software engineer in September. He is a full stack developer who brings his years of experience to maintain and expand on our mobile applications. Previously, at Rally Health, Ryan led the team building the UnitedHealthcare app from inception to release in React Native, which he now specializes in. He's also an avid birder and board gamer.


    New Board Members

    Kelly Jin

    Kelly Jin


    Kelly Jin is chief of staff to the president & CEO and vice president for AI and Insights at the John S. and James L. Knight Foundation, where she leads integration of data and responsible AI to improve strategic decision-making across programs and operations. From 2021–2025 she directed more than $200 million as head of Knight's Community and National Initiatives, supporting major civic investments across the United States. Previously she was New York City's chief analytics officer (where she launched the NYC Recovery Data Partnership and co-chaired the Algorithms Task Force) and has held roles in the White House Office of Science and Technology Policy, the Domestic Policy Council, and Boston City Hall. She holds a degree in economics from the University of Pennsylvania and first discovered iNaturalist during a trip to the Galápagos.


    Marcia Marsh

    Marcia Marsh


    Marcia was the Chief Operating Officer for the World Wildlife Fund (WWF-US). As COO, she oversaw the execution of WWF's strategy and operating systems and led government affairs and policy, private sector engagement, development and marketing and communications efforts. Before WWF, Marcia was a Price Waterhouse tax partner, Practice leader for Integrated Solutions and International Consulting at Watson Wyatt and the Vice President for Government Transformation at the Partnership for Public Service.



    Please help us welcome our new team members!

    Our staff is now seventeen people in total, and our board has nine members. We're looking forward to working on iNaturalist together with this growing team!

    iNaturalist wouldn't be possible without all the members of our amazing community — an immense thank you to each of you who has made observations, added identifications, managed taxonomy, moderated the site, translated content into multiple languages, made a donation, and so much more. Those of us on staff feel both privileged and grateful to make a living evolving the iNaturalist platform and apps, supporting the incredible community here, and furthering our mission to build a global movement for nature.

    We're also hiring a few more engineers and fundraisers — learn more here!

    Posted on November 18, 2025 04:43 PM by seastarya seastarya | 24 comments | Leave a comment

    Archives