YouTube Transcript Scraper - Captions to Notion, Slack & Jira

The YouTube captions API alternative for videos you don't own. Google's captions.download endpoint requires OAuth as the video owner; this Actor extracts the full transcript of any public YouTube video as structured JSON — every segment with start time, duration and text, in 13 languages with automatic translation fallback. No Google API key, no OAuth, no credentials stored, public captions only — with GDPR/CCPA and speaker-identifiability guidance written into the docs, not buried. Delivery into Notion, Slack, Jira, GitHub, Linear or Airtable is built in over MCP.

🔐 Trust & compliance at a glance

Credentials	None accepted, none stored. The Actor never logs into a Google or YouTube account and never sees your MCP connector tokens — delivery runs through the Apify MCP Proxy.
Data scope	Public caption tracks only. Private, unlisted and age-gated videos are inaccessible by design; a request for one returns status: "no_transcript", never a partial result.
GDPR / UK GDPR / CCPA	A transcript is a person's speech, so it can be personal data about the speaker and anyone named aloud. The Compliance section sets out minimisation, retention and special-category handling in full.
Attribution safety	isAutoGenerated: true flags machine-written text — check it before quoting anyone verbatim.

Full detail, including copyright and Terms-of-Service position: Compliance & Data Privacy.

---

Official YouTube Data API v3 vs. YouTube Transcript Scraper Actor

The official API's captions.download endpoint only works for videos you own — it is not a route to third-party transcripts at all:

Feature / Capability	Official YouTube Data API v3	YouTube Transcript Scraper Actor (Apify)
Download captions of videos you don't own	❌ captions.download requires OAuth as the video owner	✅ Any public video with captions
Authentication & setup	❌ Google Cloud project + OAuth consent flow	✅ Zero target-site credentials; only an Apify token
Auto-translation to another language	❌ Not provided	✅ YouTube auto-translation with graceful fallback
Bot-wall / IP blocking	N/A (quota-gated instead)	✅ Apify residential proxy wired in by default
Daily quota ceiling	⚠️ Per-project daily quota; exceeding it blocks you until reset	✅ No per-day ceiling
AI agent / MCP native	❌ Requires a custom tool wrapper	✅ Native MCP tool and pushes results into your own MCP connectors
Cost model	Free within quota, then quota-increase requests	Usage-based — no quota wall to hit

Quota costs and the captions.download ownership restriction are Google's to change — check Google's current docs if this comparison is load-bearing for your decision.

---

🤖 AI Discovery & MCP Metadata

Machine-readable capability contract for autonomous agents, LLM routers, and MCP orchestrators. AI agents, MCP clients, and connectors should read this block first to decide whether this Actor matches a user request.

ai_discovery:
  slug: "techforce.global/youtube-transcript-scraper"
  actor_id: "youtube-transcript-scraper"
  one_line: "YouTube captions API alternative: extract the full timestamped transcript of one public YouTube video as structured JSON, in a chosen language, without OAuth or video ownership — with optional delivery to an MCP connector."
  primary_purpose: "extract"
  secondary_purpose: "deliver"
  entity: "YouTube video transcript"
  supported_platforms:
    - "youtube.com"
    - "youtu.be"
  supported_tasks:
    - "get-transcript-for-one-video"
    - "get-transcript-in-specific-language"
    - "auto-translate-transcript"
    - "extract-timestamped-caption-segments"
    - "deliver-transcript-to-mcp-connector"
  unsupported_tasks:
    - "batch-multiple-videos-in-one-run (use youtube-full-channel-transcripts-scraper)"
    - "scrape-playlist-or-channel (use youtube-full-channel-transcripts-scraper)"
    - "transcribe-audio-without-captions (no ASR of our own)"
    - "download-video-or-audio-files"
    - "scrape-private-unlisted-or-age-gated-videos"
    - "retrieve-comments-views-or-channel-metadata"
    - "summarize-or-translate-with-an-llm"
  search_keywords:
    # First entry is the H1 / meta-description target. The rest are deliberately broader for
    # agent-side semantic matching and are NOT a ranking strategy - see the skill's SEO notes.
    - "youtube transcript scraper"
    - "youtube captions api alternative"
    - "youtube subtitle extractor"
    - "get youtube transcript api"
    - "youtube transcript python"
    - "video transcript to notion"
    - "timestamped captions json"
    - "mcp youtube tool"
    - "srt subtitle data extraction"
    - "youtube shorts transcript"
  synonyms:
    - "youtube caption downloader"
    - "youtube subtitle scraper"
    - "video transcript extractor"
    - "youtube caption json export"
  related_concepts:
    ["transcripts", "captions", "subtitles", "localization", "accessibility", "content repurposing"]
  input_entities:
    - "youtube video URL or 11-character video ID"
    - "preferred language code (1 of 13 enum values)"
    - "translate toggle"
  output_entities: ["transcript", "segment"]
  input_rules:
    videoUrl_format:
      description: "Any public YouTube URL shape, or a bare 11-character video ID. Watch links, youtu.be short links, Shorts, embed, /v/, and /live/ URLs all resolve."
      examples:
        - "https://www.youtube.com/watch?v=dQw4w9WgXcQ"
        - "https://youtu.be/dQw4w9WgXcQ"
        - "https://www.youtube.com/shorts/dQw4w9WgXcQ"
        - "dQw4w9WgXcQ"
    language_format:
      type: "enum"
      case_sensitive: false
      values: ["en", "hi", "gu", "pt", "es", "fr", "de", "it", "ja", "ko", "zh", "ru", "ar"]
  default_off_fields: []
  compliance:
    credentials_required: "none — the Actor never logs into a Google/YouTube account and stores no credential"
    data_scope: "public caption tracks only; private, unlisted and age-gated videos are inaccessible by design"
    connector_token_handling: "MCP delivery runs through the Apify MCP Proxy; the Actor never sees your connector tokens"
    personal_data_note: "a transcript is a person's speech and can be personal data about the speaker and anyone named aloud — GDPR / UK GDPR / CCPA-CPRA obligations sit with the caller"
    attribution_flag: "isAutoGenerated: true means the text is machine-written; do not attribute it as a verbatim quote"
    affiliation: "independent Actor; not affiliated with, endorsed by, or sponsored by YouTube or Google"
  limits:
    videos_per_run: 1
    dataset_items_per_run: 1
    languages_supported: 13
    requires_existing_captions: true
  pricing:
    model: "usage-based"
    billing_unit: "Apify platform usage consumed by the run"
    scope: "one video per run, regardless of transcript length"
    note: "Current rates are shown on the Actor's Apify Store page; the Console shows a live total mid-run. A hard per-run ceiling can be set with maxTotalChargeUsd."
    hard_ceiling_parameter: "maxTotalChargeUsd"
  when_to_use: >
    You need the timestamped transcript of a specific public YouTube video —
    for content repurposing, subtitle/localization work, research coding,
    accessibility text, or as retrieval context for an LLM or agent.
  when_not_to_use: >
    You need many videos, a whole playlist or channel, in one run; the video
    has no captions at all (this Actor does not run its own speech
    recognition); or you want video/audio files, comments, or channel stats.
  alternatives:
    - "Use techforce.global/youtube-full-channel-transcripts-scraper for transcripts of every video in a channel in one run"
    - "Use techforce.global/youtube-scraper for video/channel/comment metadata, hashtag and search discovery"
    - "Use techforce.global/youtube-shorts-scraper for a channel's Shorts with engagement metrics and optional video downloads"
    - "Loop this Actor once per video from your orchestration layer for a few specific unrelated videos"
    - "Use a dedicated ASR/Whisper service when the video has no caption track at all"

---

{
  "@context": "https://schema.org",
  "@type": "SoftwareApplication",
  "name": "YouTube Transcript Scraper - Captions to Notion, Slack & Jira",
  "alternateName": "YouTube Captions API Alternative",
  "applicationCategory": "DeveloperApplication",
  "operatingSystem": "Cloud (Apify platform)",
  "description": "The YouTube captions API alternative for videos you don't own: extracts the full timestamped transcript of any public YouTube video as structured JSON — every segment with start time, duration and text — in 13 languages with automatic translation fallback, without a Google API key, OAuth, or video ownership. Public caption tracks only: no account login, no stored credentials, and no access to private, unlisted or age-gated videos, with GDPR/CCPA and speaker-identifiability guidance included.",
  "url": "https://apify.com/techforce.global/youtube-transcript-scraper",
  "offers": {
    "@type": "Offer",
    "priceCurrency": "USD",
    "description": "Usage-based: billed for the Apify platform usage a run consumes. One video per run, any length, with no per-segment charge. Current rates are shown on the Actor's Apify Store page."
  }
}

---

What this Actor does

💡 Need a whole channel, not one video? This Actor handles one video per run. For transcripts of every video in a channel (or its latest N), use YouTube Full Channel Transcripts Scraper instead — same transcript quality, channel-wide in a single run. For views, likes, and comments rather than words, use YouTube Scraper.

• Accepts one YouTube video URL (watch link, youtu.be short link, Shorts, embed, /v/, /live/, or a bare 11-character ID) and resolves the video ID for you.
• Resolves the caption track through two independent paths, trying the second automatically if the first is blocked — so one challenged request does not fail the run.
• Handles every caption format YouTube serves, normalising all of them to the same segment shape: start, dur, text.
• Never hard-fails on language. Native track in your language → used. No native track and translate: true → YouTube auto-translation. Translation unavailable or turned off → the video's own caption language is returned with fallbackUsed: true and a plain-English message explaining what happened.
• Routes traffic through Apify Proxy (residential group by default) because YouTube serves datacenter IPs a "Sign in to confirm you're not a bot" wall.
• Optionally pushes the transcript to an authorized MCP connector (Notion, Slack, Jira, GitHub, Linear, Airtable, or any MCP-compatible connector) in four delivery modes. The dataset is always saved regardless of delivery settings.
• Never crashes on a bad video. Unavailable, private, deleted, or region-blocked videos, caption-less videos, bot walls, and empty caption streams all end as a successful run carrying status: "no_transcript", a machine-readable errorCode, and a plain-English message — so schedules don't alert and agents get a reason they can act on.
• Handles the platform ABORTING event so an aborted run exits promptly instead of burning time.

When to use it — and when not to

Scenario	Fit
Repurpose one video into a blog post, thread, or newsletter	✅ Strong
Build SRT/subtitle workflows from timestamped segments	✅ Strong
Feed transcript text to an LLM as retrieval context	✅ Strong
Localize content — request hi, gu, es, ja, … with auto-translation	✅ Strong
Accessibility text alternatives for audio/video	✅ Strong
Push a transcript straight into Notion or Slack	✅ Strong (native MCP delivery)
Research coding / qualitative analysis of a specific talk	✅ Strong
Every video in a channel	⚠️ One video per run — use YouTube Full Channel Transcripts Scraper
A handful of specific, unrelated videos	⚠️ Loop the Actor — one run per video
A language outside the 13-value enum	⚠️ Not selectable — request en and translate downstream
Video with no caption track at all	❌ No transcript — the run still succeeds, returning status: "no_transcript" and the reason
Private, unlisted, or age-gated videos	❌ Not supported — public videos only
Views, likes, comments, or channel stats	❌ Out of scope — use YouTube Scraper
A channel's Shorts catalogue or video file downloads	❌ Out of scope — use YouTube Shorts Scraper
Summarization or LLM translation of the transcript	❌ Not performed — returns raw caption text

---

⚡ Execution Rules (Avoid Common Agent Failures)

1. videoUrl is required and has no default. Unlike some Actors, omitting it — or passing something with no recognisable video ID — no longer fails the run: the Actor reports errorCode: "MISSING_VIDEO_URL" or "BAD_VIDEO_URL" on a SUCCEEDED run with one dataset item explaining what to pass. It does not silently scrape a sample video either. The Console pre-fills a demo URL; API/MCP callers must pass their own.
2. A bare 11-character video ID is accepted. dQw4w9WgXcQ works exactly like a full URL. As a last resort the Actor also searches for any 11-character token in the string, so a malformed input can resolve to an unintended video — always pass a clean URL or ID and check the returned videoId.
3. language is a fixed 13-value enum, lowercase ISO-639-1: en, hi, gu, pt, es, fr, de, it, ja, ko, zh, ru, ar. There is no en-US, no pt-BR, and no free-text language input.
4. The language you ask for is not guaranteed. Read language, isTranslated, fallbackUsed, and message on the output item — not just your own request. A successful run can legitimately return a different language than requested.
5. translate: false does not fail the run. It disables auto-translation; the Actor then returns the video's available caption language with fallbackUsed: true and a message listing the available languages. Branch on fallbackUsed, not on run status.
6. Captions must already exist. This Actor reads YouTube's caption tracks; it performs no speech recognition. A video with no captions still produces a successful run carrying status: "no_transcript" and errorCode: "NO_CAPTIONS" — not a crash. 6a. Video-state problems never fail the run. Unavailable, private, deleted, or region-blocked videos, bot walls, and empty caption streams all return SUCCEEDED with exactly one dataset item whose status is "no_transcript", plus an errorCode and a plain-English message. Always branch on item["status"], not on run status — a SUCCEEDED run does not guarantee a transcript. A missing or unparseable videoUrl is reported the same way, as MISSING_VIDEO_URL or BAD_VIDEO_URL — no input mistake fails the run.
7. Keep the residential proxy on. The default is { useApifyProxy: true, apifyProxyGroups: ["RESIDENTIAL"] }. Datacenter IPs regularly hit YouTube's bot wall, which surfaces as a caption-retrieval failure, not a proxy error. Residential groups require a paid Apify plan.
8. One run = one video = one dataset item. The transcript is a nested searchResult array inside that single item, not one item per segment. itemCount: 1 is a full result.
9. start and dur are strings, not numbers — e.g. "3.200". Cast before doing arithmetic.
10. deliveryMode defaults to summary, not none. For dataset-only runs, either leave mcpConnector empty or set deliveryMode: "none" explicitly.
11. perSegment delivery issues one connector call per segment. A 20-minute video can be hundreds of calls — prefer chunked or summary unless you specifically want per-line records.

dependencies:
  - rule: "delivery requires mcpConnector AND mcpTool"
    on_violation: "delivery step is skipped with a warning; the dataset item is still written"
  - rule: "mcpArguments must reference {message} for the rendered template to appear anywhere"
    on_violation: "message body is empty; the tool is still called"
  - rule: "mcpConnector only resolves when the Actor runs on the Apify platform"
    on_violation: "local runs log 'APIFY_MCP_PROXY_URL is not set' and skip delivery; the dataset is unaffected"
  - rule: "chunked mode exposes {part} and {partCount}"
    note: "put them in the page/message title or every part looks identical"
authentication:
  actor_input: none
  platform: "Apify API token required for API / SDK / MCP invocation; paid plan required for residential proxy groups"
  target_site: "no YouTube or Google account, API key, or OAuth app required (public captions only)"

---

📥 Input Contract

.actor/input_schema.json is authoritative — if any table on this page disagrees with the schema files, the schema files win. API and MCP clients should always pass videoUrl explicitly.

Scraping Parameters

Field	Type	Required	Default	Allowed Values / Format	Example
videoUrl	string	✅ Yes	none (Console prefill only)	Any public YouTube URL shape, or a bare 11-char ID	https://youtu.be/dQw4w9WgXcQ
language	string (enum)	⬜ No	en	13 lowercase codes (see below)	hi
translate	boolean	⬜ No	true	true / false	false
proxyConfiguration	object	⬜ No	{ useApifyProxy: true, apifyProxyGroups: ["RESIDENTIAL"] }	Apify proxy editor object	see below

Accepted URL shapes

Shape	Example
Standard watch	https://www.youtube.com/watch?v=dQw4w9WgXcQ
Short link	https://youtu.be/dQw4w9WgXcQ
Shorts	https://www.youtube.com/shorts/dQw4w9WgXcQ
Embed	https://www.youtube.com/embed/dQw4w9WgXcQ
Live	https://www.youtube.com/live/dQw4w9WgXcQ
Mobile / music / nocookie hosts	https://m.youtube.com/watch?v=dQw4w9WgXcQ
Bare video ID	dQw4w9WgXcQ

Language reference — all 13 values

en English · hi हिन्दी / Hindi · gu ગુજરાતી / Gujarati · pt Português · es Español · fr Français · de Deutsch · it Italiano · ja 日本語 · ko 한국어 · zh 中文 · ru Русский · ar العربية

Region variants (en-US, pt-BR, zh-Hans) are not valid inputs — pass the base code.

Language resolution order

Situation	What you get	isTranslated	fallbackUsed	message
Native caption exists in requested language	That track	false	false	null
No native track, translate: true	YouTube auto-translation to your language	true	false	explains the translation
Translation returns no text	Video's own caption language	false	true	explains the failed translation
No native track, translate: false	Video's own caption language	false	true	lists the available languages
No language passed at all	Video's default caption track	false	false	null
No caption tracks at all	status: "no_transcript", errorCode: "NO_CAPTIONS" — run still succeeds	false	false	explains that the video has no captions

MCP Delivery Configuration (Optional)

Parameter	Type	Default	Description
mcpConnector	string	""	Authorized MCP connector on your Apify account (Notion, Slack, Jira, GitHub, Linear, Airtable, …). Leave empty for dataset-only runs.
deliveryMode	enum	summary	summary (one call, full transcript) · chunked (split long transcripts) · perSegment (one call per segment) · none.
mcpTool	string	""	Tool name on the connector: create_page, send_message, create_issue, append_row. Run once with a connector selected — the log lists the connector's available tools.
mcpArguments	object	{}	Arguments passed to the tool. String leaves support {placeholders}.
mcpMessageTemplate	string	""	Template rendered and exposed to mcpArguments as {message}.

Placeholders by mode:

Mode	Placeholders
summary	{videoUrl}, {videoId}, {language}, {segmentCount}, {transcript} (full text), {message}
chunked	the above, plus {part} and {partCount} — and {transcript} holds one part
perSegment	{start}, {dur}, {text}, plus {videoUrl}, {videoId}, {language}, {segmentCount}, {message}

Delivery packs segments into paragraphs of ~1,800 characters and groups them into parts under ~72,000 characters, so services with per-request block or timeout caps (Notion in particular) never reject the call.

{
  "videoUrl": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
  "language": "en",
  "translate": true,
  "mcpConnector": "<your-authorized-slack-connector>",
  "deliveryMode": "summary",
  "mcpTool": "send_message",
  "mcpMessageTemplate": "Transcript of {videoUrl} ({language}, {segmentCount} segments):\n\n{transcript}",
  "mcpArguments": { "channel": "#transcripts", "text": "{message}" }
}

Notion, chunked across pages (best for long videos):

{
  "videoUrl": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
  "language": "hi",
  "mcpConnector": "<your-authorized-notion-connector>",
  "deliveryMode": "chunked",
  "mcpTool": "notion-create-pages",
  "mcpArguments": {
    "parent": { "page_id": "<your-page-id>" },
    "pages": [
      {
        "properties": { "title": "Transcript {videoId} (part {part}/{partCount})" },
        "content": "{transcript}"
      }
    ]
  }
}

---

📤 Output Contract & Data Structure

Storage: Apify Dataset — exactly one item per run, with the transcript nested in searchResult. Pagination: limit & offset on the dataset items endpoint (rarely needed for a single item). Duplicates: none within a run. Across runs, dedupe on videoId + language.

Every field below is always present on the item — the shape never changes. What varies is the value, and the Toggle / controlled by column names the input that decides it. No output field is toggled on or off by configuration.

Field	Type	Nullable	Toggle / controlled by	Meaning
videoUrl	string	No	videoUrl	Normalized https://www.youtube.com/watch?v=<id> URL — not necessarily the URL you passed
videoId	string	No	videoUrl	Extracted 11-character video ID — verify this matches the video you meant
requestedLanguage	string	No (may be "")	language	The language code you asked for, trimmed — echoed back verbatim
language	string	No (may be "")	language + translate	The language actually returned. "" when no transcript was found
isAutoGenerated	boolean	No	— (source track)	true when the source track is YouTube's ASR (auto-generated) captions
isTranslated	boolean	No	translate	true when the text is a YouTube auto-translation
fallbackUsed	boolean	No	language + translate	true when the requested language could not be honoured
message	string	Yes (null)	language + translate	Plain-English note about any language/translation fallback, or why no transcript exists; null on a clean match
segmentCount	integer	No	—	Number of segments in searchResult; 0 when status is "no_transcript"
searchResult	array	No (may be [])	—	Transcript segments, in playback order; [] when status is "no_transcript"
searchResult[].start	string	No (may be "")	—	Segment start in seconds, as a string — e.g. "3.200"
searchResult[].dur	string	No (may be "")	—	Segment duration in seconds, as a string — e.g. "2.900"
searchResult[].text	string	No	—	Segment text, HTML-unescaped, newlines collapsed to spaces. Empty segments are dropped.
status	string	No	— (video state)	"ok" when a transcript was extracted, "no_transcript" when none could be — branch on this first
errorCode	string	Yes (null)	— (video state)	null when status is "ok"; otherwise the machine-readable reason (see the matrix below)

Note for consumers: message is null (not "") when nothing needed explaining, and start/dur are strings for compatibility with the widely-used searchResult transcript format. Cast them before arithmetic, and test message for None/null.

Two item shapes. Every run writes exactly one item, but it comes in two shapes distinguished by status:

status	errorCode	segmentCount	searchResult	message
"ok"	null	≥ 1	populated	null, or a language-fallback note
"no_transcript"	one of MISSING_VIDEO_URL · BAD_VIDEO_URL · VIDEO_UNAVAILABLE · NO_CAPTIONS · CAPTIONS_BLOCKED · CAPTIONS_UNREACHABLE · EMPTY_TRANSCRIPT · UNEXPECTED_ERROR	0	[]	always populated — the human-readable reason

Example output item

{
  "videoUrl": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
  "videoId": "dQw4w9WgXcQ",
  "requestedLanguage": "en",
  "language": "en",
  "isAutoGenerated": false,
  "isTranslated": false,
  "fallbackUsed": false,
  "message": null,
  "segmentCount": 3,
  "searchResult": [
    { "start": "0.000", "dur": "3.200", "text": "Welcome back to the channel." },
    { "start": "3.200", "dur": "2.900", "text": "Today we're talking about web scraping." },
    { "start": "6.100", "dur": "4.000", "text": "Let's jump right into the setup." }
  ],
  "status": "ok",
  "errorCode": null
}

Example "no transcript" item

An unavailable, private, deleted, or caption-less video produces a successful run with this shape instead:

{
  "videoUrl": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
  "videoId": "dQw4w9WgXcQ",
  "requestedLanguage": "en",
  "language": "",
  "isAutoGenerated": false,
  "isTranslated": false,
  "fallbackUsed": false,
  "message": "This video is not available on YouTube (YouTube says: \"Video unavailable\"). It may be private, deleted, region-blocked, or the ID may be wrong, so there is no transcript to extract.",
  "segmentCount": 0,
  "searchResult": [],
  "status": "no_transcript",
  "errorCode": "VIDEO_UNAVAILABLE"
}

Fallback example — Hindi requested, only English captions exist, translate: false:

{
  "requestedLanguage": "hi",
  "language": "en",
  "isTranslated": false,
  "fallbackUsed": true,
  "message": "The requested language 'hi' is not available and translation is turned off. Returning the available caption language 'en'. Available languages: en, es."
}

What a successful run looks like

{
  "status": "SUCCEEDED",
  "defaultDatasetId": "<datasetId>",
  "stats": { "itemCount": 1 }
}

itemCount: 1 is the complete result — but it does not by itself mean a transcript was extracted. Read the item's status:

• status: "ok" → transcript present; segmentCount gives its length.
• status: "no_transcript" → the run worked, the video didn't yield captions. Read errorCode and message for the reason, and the run's terminal status message shows No transcript (<CODE>): <message>.

Nothing about the input or the video fails the run any more. A missing or unparseable videoUrl, and every fact about the video's own state, is reported as a SUCCEEDED run carrying an errorCode and a plain-English message. A FAILED run now indicates a platform-level problem — an aborted run, a timeout, or an exhausted spend cap — not something your input said.

---

▶️ Invocation & Integration

Lifecycle: Discover → Understand → Configure → Execute → Retrieve → Validate. Authenticate every call with Authorization: Bearer <APIFY_TOKEN>.

Apify Console

Open the Actor → Input tab → paste the YouTube video URL → pick a Preferred language → leave Translate on → (optional) choose an MCP connector, tool, and template → Start → read the Dataset tab (JSON, CSV, Excel, XML).

Model Context Protocol (MCP)

Add this Actor directly as an AI tool in Claude Desktop, Cursor, VS Code, or LangChain agents:

claude mcp add --transport http apify \
  "https://mcp.apify.com?tools=techforce.global/youtube-transcript-scraper"

Python (apify-client)

import os
from apify_client import ApifyClient

client = ApifyClient(os.getenv("APIFY_TOKEN"))

run = client.actor("techforce.global/youtube-transcript-scraper").call(run_input={
    "videoUrl": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
    "language": "en",
    "translate": True,
    "deliveryMode": "none",
})

item = client.dataset(run["defaultDatasetId"]).list_items().items[0]

# Always branch on status: a SUCCEEDED run can legitimately carry no transcript.
if item["status"] != "ok":
    raise SystemExit(f'No transcript ({item["errorCode"]}): {item["message"]}')

print(item["language"], item["segmentCount"], "segments")
if item["fallbackUsed"]:
    print("Language fallback:", item["message"])

full_text = " ".join(seg["text"] for seg in item["searchResult"])
print(full_text[:400])

Batch several videos by looping the Actor (one video per run):

urls = ["https://youtu.be/aaaaaaaaaaa", "https://youtu.be/bbbbbbbbbbb"]
for url in urls:
    run = client.actor("techforce.global/youtube-transcript-scraper").call(
        run_input={"videoUrl": url, "language": "en", "deliveryMode": "none"}
    )
    print(url, run["status"])

JavaScript / TypeScript (apify-client)

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: process.env.APIFY_TOKEN });

const run = await client.actor('techforce.global/youtube-transcript-scraper').call({
  videoUrl: 'https://youtu.be/dQw4w9WgXcQ',
  language: 'hi',
  translate: true,
  deliveryMode: 'none',
});

const { items } = await client.dataset(run.defaultDatasetId).listItems();
const [transcript] = items;

if (transcript.status !== 'ok') {
  // Unavailable video, no captions, bot wall — the run succeeded, the video didn't.
  throw new Error(`No transcript (${transcript.errorCode}): ${transcript.message}`);
}

console.log(`${transcript.language}: ${transcript.segmentCount} segments`);
console.log(transcript.searchResult.map((s) => s.text).join(' ').slice(0, 400));

cURL — synchronous (returns the item directly; 300s limit)

curl -X POST \
  -H 'Content-Type: application/json' \
  -H 'Authorization: Bearer <YOUR_APIFY_TOKEN>' \
  -d '{"videoUrl":"https://www.youtube.com/watch?v=dQw4w9WgXcQ","language":"en","deliveryMode":"none"}' \
  'https://api.apify.com/v2/acts/techforce.global~youtube-transcript-scraper/run-sync-get-dataset-items'

cURL — asynchronous with a spend cap (recommended for production)

# 1. Start execution with a hard spend limit
curl -X POST -H 'Content-Type: application/json' \
  -H 'Authorization: Bearer <YOUR_APIFY_TOKEN>' \
  -d '{"videoUrl":"https://youtu.be/dQw4w9WgXcQ","language":"en","deliveryMode":"none"}' \
  'https://api.apify.com/v2/acts/techforce.global~youtube-transcript-scraper/runs?maxTotalChargeUsd=1'

# 2. Poll
curl -H 'Authorization: Bearer <YOUR_APIFY_TOKEN>' \
  'https://api.apify.com/v2/actor-runs/<runId>'

# 3. Retrieve when SUCCEEDED
curl -H 'Authorization: Bearer <YOUR_APIFY_TOKEN>' \
  'https://api.apify.com/v2/datasets/<DATASET_ID>/items?clean=true&format=json'

---

⚠️ Autonomous Agent Error Handling & Resolution Matrix

Read this first: every problem with the video is reported as a SUCCEEDED run whose single dataset item carries status: "no_transcript" and an errorCode. Caller mistakes are reported identically: a missing videoUrl as MISSING_VIDEO_URL, an unusable one as BAD_VIDEO_URL. No input error fails the run. An agent that checks run status alone will mistake "this video has no captions" for success with data; check item["status"]. Connector delivery is skipped entirely when there is no transcript, so a no_transcript outcome never posts an empty message to Slack or Notion.

Error Code	Detection Condition	Underlying Cause	Deterministic Agent Action
AUTH_INVALID	HTTP 401	Missing or malformed APIFY_TOKEN	Abort & Prompt User: request a valid API token.
RATE_LIMITED	HTTP 429	Apify API rate / concurrency limits	Retry: exponential backoff (2s, 4s, 8s).
SYNC_TIMEOUT	HTTP 408 on the sync endpoint	Run exceeded the 300s synchronous ceiling	Reconfigure: switch to async POST /runs + polling.
MISSING_VIDEO_URL	SUCCEEDED, item errorCode: "MISSING_VIDEO_URL", status: "no_transcript"	videoUrl omitted or blank (no default exists)	Modify Input: pass videoUrl explicitly. The message field says what shapes are accepted.
BAD_VIDEO_URL	SUCCEEDED, item errorCode: "BAD_VIDEO_URL", status: "no_transcript"; the offending value is echoed in videoUrl	Not a YouTube URL / no 11-char ID present	Modify Input: pass a clean watch URL, youtu.be link, or bare ID.
WRONG_VIDEO	SUCCEEDED but videoId ≠ intended video	Last-resort 11-char token match latched onto the wrong substring	Validate & Re-run: always compare returned videoId against your input.
VIDEO_UNAVAILABLE	SUCCEEDED, item errorCode: "VIDEO_UNAVAILABLE"	Video is private, deleted, region-blocked, or the ID is wrong — both fetch paths report it unavailable	Do not retry. Verify the video opens in a browser; ask the user for a valid URL.
NO_CAPTIONS	SUCCEEDED, item errorCode: "NO_CAPTIONS"	Video exists but has no caption tracks at all	Do not retry. Route to an ASR/Whisper service; this Actor cannot transcribe audio.
CAPTIONS_BLOCKED	SUCCEEDED, item errorCode: "CAPTIONS_BLOCKED"	YouTube challenged the request (bot wall, sign-in, age gate)	Reconfigure & retry once: enable the RESIDENTIAL proxy group.
CAPTIONS_UNREACHABLE	SUCCEEDED, item errorCode: "CAPTIONS_UNREACHABLE"	Neither resolution path returned usable caption metadata	Retry once with a residential proxy; if it repeats, treat as unavailable.
EMPTY_TRANSCRIPT	SUCCEEDED, item errorCode: "EMPTY_TRANSCRIPT"	Caption track resolved but parsed to zero usable text	Retry: re-run with translate: true or a different language.
UNEXPECTED_ERROR	SUCCEEDED, item errorCode: "UNEXPECTED_ERROR"	An unhandled upstream shape or network fault; the log carries the stack trace	Retry once, then report the run ID.
PROXY_UNAVAILABLE	Log Could not set up proxy configuration or Running without a proxy	Residential groups require a paid Apify plan	Prompt User: upgrade the plan; expect bot-wall failures until then.
LANGUAGE_FALLBACK	SUCCEEDED with fallbackUsed: true	Requested language unavailable	Not an error. Read language and message; re-run with translate: true if a translation is acceptable.
TRANSLATION_FALLBACK	SUCCEEDED with isTranslated: false after requesting a foreign language	Translation attempt returned no text	Not an error. Consume the returned language or translate downstream.
DELIVERY_SKIPPED	Log no tool name was provided	mcpConnector set but mcpTool empty	Modify Input: set mcpTool; the dataset item is already saved.
DELIVERY_TOOL_UNKNOWN	Log Tool 'x' is not available on this connector	Wrong tool name for that connector	Modify Input: pick a name from the Available tools: list in the same log line.
DELIVERY_EMPTY	Connector called, body empty	mcpArguments omitted the {message} placeholder	Modify Input: map {message} inside mcpArguments.
MCP_PROXY_MISSING	Log APIFY_MCP_PROXY_URL is not set	Running locally instead of on the platform	Reconfigure: apify push and run on the platform.
CHARGE_LIMIT	Run aborted immediately, before any work	maxTotalChargeUsd set below what the run costs to start	Prompt User: raise the cap. Check the Store page for the current rate.

Provenance of this table: every errorCode, status value and log-line condition above is taken from this Actor's own source and verified against it. The four HTTP rows (401, 408, 429) describe the Apify API transport layer, not this Actor — confirm them against the Apify API reference before wiring them into automated retry logic.

---

🗣️ Natural Language → Actor Mapping

User says	Intent	Constructed Actor input
"Get me the transcript of this YouTube video"	Transcript extraction	{"videoUrl":"<url>","language":"en"}
"What is said in this video?"	Content Q&A / retrieval context	{"videoUrl":"<url>"} then read searchResult[].text
"Summarize this YouTube video"	Summarization input	{"videoUrl":"<url>"} → summarize searchResult text yourself (the Actor returns raw captions)
"Give me the Hindi transcript of this talk"	Localized transcript	{"videoUrl":"<url>","language":"hi","translate":true}
"Transcript in Spanish, but only if it's a real Spanish caption"	Strict language	{"videoUrl":"<url>","language":"es","translate":false} then check fallbackUsed
"Pull the captions of this Short"	Shorts extraction	{"videoUrl":"https://www.youtube.com/shorts/<id>"}
"I need timestamps so I can cut clips"	Subtitle / clipping workflow	{"videoUrl":"<url>"} then read start + dur per segment
"Turn this video into SRT subtitles"	Localization	{"videoUrl":"<url>","language":"en"} then format segments as SRT downstream
"Save this video's transcript to Notion"	Pipeline/delivery	{"videoUrl":"<url>","mcpConnector":"notion","deliveryMode":"chunked","mcpTool":"notion-create-pages"}
"Post the transcript to Slack #content"	Pipeline/delivery	{"videoUrl":"<url>","mcpConnector":"slack","deliveryMode":"summary","mcpTool":"send_message","mcpArguments":{"channel":"#content","text":"{message}"}}
"Open a Jira ticket with this transcript attached"	Pipeline/delivery	{"videoUrl":"<url>","mcpConnector":"jira","deliveryMode":"summary","mcpTool":"create_issue"}
"Get transcripts for these 12 videos"	Batch extraction	12 sequential runs, one videoUrl each — then merge on videoId
"Get transcripts for every video on this channel"	Channel-wide extraction	➡️ Route to YouTube Full Channel Transcripts Scraper
"How many views and comments does this video have?"	Metadata	➡️ Route to YouTube Scraper
"List all the Shorts on this channel"	Shorts catalogue	➡️ Route to YouTube Shorts Scraper
"Transcribe this video — it has no captions"	ASR request	❌ Out of scope — route to a speech-recognition service

Should NOT route here: videos with no caption track (needs ASR) · playlists, channels, or batch lists in a single run (→ Full Channel Transcripts) · comments, views, likes, or channel statistics (→ YouTube Scraper) · a channel's Shorts or video-file downloads (→ Shorts Scraper) · private, unlisted, or age-gated videos · LLM summarization or translation of the text · non-YouTube video platforms (Vimeo, TikTok, Twitch).

---

🧭 Agent Execution & Routing Logic

[Input User Query]
       │
       ▼
1. Target is a YouTube video (youtube.com / youtu.be)?  NO  → Route to the correct platform Actor
       │ YES
       ▼
2. Wants transcript/caption text (not audio, stats, comments)?  NO  → Route elsewhere / abort
       │ YES
       ▼
3. Is the video public (not private/unlisted/age-gated)?        NO  → Reject
       │ YES
       ▼
4. Does the video have captions?                    UNKNOWN → Try; if errorCode == NO_CAPTIONS route to an ASR service
       │
       ▼
5. Can a video ID be extracted from the input?      NO  → Prompt user for a clean URL or 11-char ID
       │ YES
       ▼
6. Is the requested language in the 13-value enum?  NO  → Use "en" and translate downstream
       │ YES
       ▼
7. Is a non-requested language acceptable?          NO  → set translate:false and check fallbackUsed
       │ YES → keep translate:true (default)
       ▼
8. More than one video needed?                      YES → Whole channel? Route to Full Channel Transcripts Scraper
       │ NO                                                 A few videos? Loop: one run per video
       ▼
9. Deliver to a connector?                          YES → set mcpConnector + mcpTool + {message} in mcpArguments
       │ NO  → set deliveryMode "none"
       ▼
[Execute Apify Actor]
       │
       ├──► run "SUCCEEDED" ──► item.status == "ok"?
       │                          ├── YES → verify videoId, read language / isTranslated / fallbackUsed / message → consume searchResult
       │                          └── NO  → read item.errorCode → Error Handling Matrix (retry, re-route to ASR, or ask the user)
       └──► run "FAILED"    ──► platform-level only (abort, timeout, spend cap) → not an input problem

---

💰 Cost Control & Pricing Transparency

Billing model: usage-based. You pay for the Apify platform usage a run consumes. Current rates are on the Actor's Apify Store page, and the Console shows a live total while a run is in progress.

• One run = one video, any length. There is no per-segment or per-character charge, so a 3-minute clip and a 3-hour lecture are the same shape of run.
• Transcript retrieval is short and bounded. The expensive variable is not video length — it is whether the request is challenged and has to fall back.
• Autonomous spend caps: pass maxTotalChargeUsd on the run endpoint for a hard per-execution ceiling. It is the only hard stop; input validation bounds the work, not the spend.

Worked examples — relative run shape

Multiply by your account's usage rate from the Store page to get a currency figure.

Scenario	Runs	Relative cost per run	Notes
One video, captions present	1	Baseline	The normal case.
One video, 3-hour lecture	1	≈ Baseline	Length barely matters — the caption track is one fetch.
One video, primary path challenged	1	Slightly above baseline	Fallback path adds one request.
One video, no captions at all	1	Below baseline	Ends early with NO_CAPTIONS; no parsing work.
Same video in 3 languages	3	3 × baseline	Each language is a separate run.
50 unrelated videos	50	50 × baseline	Loop from your orchestrator.
A whole channel	1	—	➡️ Use Full Channel Transcripts Scraper instead of 200 runs here.

Cost-control tips: validate the URL before spending a run; check the video's transcript panel in a browser first if you are unsure captions exist; cache results by videoId + language so a repeat request costs nothing; and remember each language is a separate run — so for multi-language work, decide the set up front.

---

🛠️ Troubleshooting

Most failures are caption availability or proxy issues, not scraper bugs. Work through this table top to bottom — the first three rows account for the large majority of support tickets.

#	Symptom you see	Most likely cause	Fix
1	SUCCEEDED but errorCode: "VIDEO_UNAVAILABLE" — This video is not available on YouTube…	The video is private, deleted, region-blocked, or the ID is wrong. Both fetch paths agreed	Open the URL in a browser. If it doesn't play for you either, no scraper can reach it — retrying and changing proxy will not help.
2	SUCCEEDED but errorCode: "CAPTIONS_BLOCKED" / "CAPTIONS_UNREACHABLE"	YouTube challenged the request, or returned no caption metadata on either path	Enable RESIDENTIAL Apify Proxy (the default) and retry once. Datacenter/local IPs are blocked routinely.
3	SUCCEEDED but errorCode: "NO_CAPTIONS"	The video genuinely has no caption track	Expected. This Actor reads captions; it does not transcribe audio. Use an ASR/Whisper service.
3a	SUCCEEDED but errorCode: "EMPTY_TRANSCRIPT"	Caption track resolved but every segment parsed empty	Re-run with translate: true, or request a different language.
3b	Your pipeline treats a no_transcript run as a transcript	You checked run status instead of item["status"]	Branch on item["status"] == "ok" before reading searchResult.
4	Wrong language in the output	Requested language had no native track; you got a translation or a fallback	Read language, isTranslated, fallbackUsed, message. Set translate: false if only a native track will do.
5	translate: false but the run still succeeded in another language	By design — it disables translation, it does not fail the run	Branch on fallbackUsed: true, not on run status.
6	en-US / pt-BR rejected or ignored	Only the 13 base enum codes are valid	Pass en, pt, … Region variants are not selectable.
7	Text reads like speech-to-text with no punctuation	The source track is YouTube's auto-generated ASR caption	Check isAutoGenerated: true. Nothing to fix — that is the only track available.
8	videoId is not the video you meant	A last-resort regex matched an 11-char token in a malformed input	Always pass a clean URL or bare ID, and assert the returned videoId.
9	Only one dataset item	By design — one run = one video = one item, transcript nested in searchResult	Read segmentCount for length; unwind searchResult for a per-segment table.
10	start / dur arithmetic fails	They are strings ("3.200"), not numbers	Cast to float before arithmetic.
11	message is null in your consumer	null means "nothing needed explaining"	Test for None/null, not for "".
12	Log: Running without a proxy	No proxy configured or unavailable on the plan	Enable Apify Proxy residential group; a paid plan is required.
13	Log shows the primary resolution path failing, but the run succeeded	Normal — the fallback path took over	No action needed.
14	Need many videos and each run is a separate charge	One video per run by design	For a whole channel use Full Channel Transcripts Scraper; otherwise loop and cache by videoId + language.
15	Slack / Notion / Jira received nothing	Delivery needs both mcpConnector and mcpTool; otherwise it is skipped with a warning	Set both. The dataset item is still saved — your data is never lost.
16	Log: APIFY_MCP_PROXY_URL is not set	You ran locally with apify run; connectors only resolve on the platform	Deploy with apify push and run on the platform.
17	Connector fired, but the message body is empty	mcpArguments did not reference {message}	Map {message} in mcpArguments and populate mcpMessageTemplate.
18	Notion rejects or truncates a long transcript	Single-call payload exceeded block/timeout caps	Use deliveryMode: "chunked" and put {part}/{partCount} in the page title.
19	Every Notion page looks identical	{part}/{partCount} missing from the title template	Add them: "Transcript {videoId} (part {part}/{partCount})".
20	Hundreds of connector calls fired	deliveryMode: "perSegment" sends one call per segment	Switch to chunked or summary unless per-line records are the goal.
21	Delivery fired when you only wanted a dataset	deliveryMode defaults to summary, not none	Set deliveryMode: "none" or leave mcpConnector empty.
22	HTTP 401 / 403	Missing, expired, or malformed APIFY_TOKEN	Regenerate in Apify Console → Settings → API & Integrations, send as Authorization: Bearer <token>.
23	HTTP 429	Apify account concurrency / rate limits — not YouTube blocking	Retry with exponential backoff (2s → 4s → 8s).
24	HTTP 408 on run-sync-get-dataset-items	The synchronous endpoint has a hard 300-second ceiling	Switch to async: POST /runs → poll /actor-runs/<runId> → fetch dataset.

Diagnostic checklist before opening an issue

1. Open the video in a browser and check the transcript panel (⋯ → Show transcript). If YouTube shows nothing there, the video has no captions and no scraper can produce them.
2. Confirm the proxy. Residential group on, paid plan active. This is the single most common cause of caption-retrieval failures.
3. Baseline run. videoUrl only, language: "en", translate: true, deliveryMode: "none".
4. Check the resolved videoId in the output or the run log against the video you meant.
5. Read the run log — it names which resolution path succeeded and which caption language was selected.

If the issue survives all five steps, open an Issues ticket on the Actor page (or email support) with the run ID, the video URL, and what you expected. Run IDs let us replay the exact execution.

❓ FAQ

Setup & access

Do I need a YouTube or Google account, API key, or OAuth app?

No. The Actor reads publicly available caption tracks. You need an Apify account and API token; you need no Google credentials whatsoever.

Is this the official YouTube API?

No. This is an independent Actor and is not affiliated with, endorsed by, or sponsored by YouTube or Google.

Why can't I just use the official YouTube Data API?

captions.download only works for videos you own, via OAuth. For any third-party video it is not an option — which is the main reason this Actor exists.

Do I need to configure proxies?

The residential Apify Proxy is on by default and you should leave it on — YouTube serves datacenter IPs a bot-check wall. Residential groups require a paid Apify plan.

Can I plug it into Claude, Cursor, or a LangChain agent?

Yes — it is a native MCP tool:

claude mcp add --transport http apify "https://mcp.apify.com?tools=techforce.global/youtube-transcript-scraper"

Input & languages

Can I pass a playlist, channel, or list of videos?

No — one video per run. For a whole channel, use YouTube Full Channel Transcripts Scraper, which takes a channel URL and processes up to 50 videos concurrently. For a few specific unrelated videos, loop this Actor from your orchestration layer (each video is a separate run, and a separate charge).

Which languages are supported?

13: en, hi, gu, pt, es, fr, de, it, ja, ko, zh, ru, ar. Region variants like en-US are not valid inputs.

What happens if my language isn't available?

With translate: true (default) you get a YouTube auto-translation. With translate: false you get the video's own caption language, fallbackUsed: true, and a message listing what was available. The run does not fail for language reasons.

Does it work on Shorts?

Yes — Shorts, embed, /live/, youtu.be, mobile and youtube-nocookie hosts, and bare 11-character IDs all resolve.

Output & data

Why is the whole transcript in one dataset item?

One run handles one video, so the transcript is a nested searchResult array on a single item. Use the Console's Transcript view (or unwind=searchResult on the API) for a per-segment table.

Why are start and dur strings?

For compatibility with the widely-used searchResult transcript format. Cast them to float before doing arithmetic.

Does it summarize or translate with an LLM?

No. It returns YouTube's caption text — raw, timestamped, and unmodified apart from HTML unescaping. Summarize downstream.

Can it transcribe a video that has no captions?

No. There is no speech recognition in this Actor. Use an ASR service for caption-less videos. The run itself still succeeds and tells you so, with status: "no_transcript" and errorCode: "NO_CAPTIONS".

Why did my run succeed but return no transcript?

Because the video, not the Actor, was the problem: it may be private, deleted, region-blocked, caption-less, or YouTube may have challenged the request. Rather than crashing, the Actor writes one dataset item with status: "no_transcript", a machine-readable errorCode, and a plain-English message — so scheduled runs and agent pipelines get a readable reason instead of an alert storm from failed runs. Read item["status"] first, always.

Which outcomes still fail the run?

None of them. A missing videoUrl (MISSING_VIDEO_URL), an unusable one (BAD_VIDEO_URL) and every fact about the video's own state all end as a SUCCEEDED run whose single item carries an errorCode and a validation message — so API and MCP callers get a readable reason instead of a stack trace. A FAILED run now means a platform-level problem: an abort, a timeout, or an exhausted spend cap.

Delivery

What is the MCP connector for?

It pushes the transcript straight into a tool you have already authorized on your Apify account — Notion, Slack, Jira, GitHub, Linear, Airtable, or any MCP-compatible connector — with no glue code. Credentials stay private: delivery runs through the Apify MCP Proxy, so the Actor never sees your connector tokens.

Which delivery mode should I use?

summary for short videos, chunked for long transcripts into services with request caps (Notion), perSegment for per-line records (logs, row-per-segment sheets), none for dataset-only runs.

Does delivery work when running locally?

No. Connectors resolve through the Apify MCP Proxy, which only exists on the platform. Locally you'll see APIFY_MCP_PROXY_URL is not set and the dataset item is saved as normal.

Does it deliver anything when there is no transcript?

No. A no_transcript outcome skips delivery entirely, so your Slack channel or Notion database never receives an empty transcript. The reason is in the dataset item and the run's status message.

Limits & pricing

What does it cost?

Usage-based: you pay for the platform usage a run consumes. One video per run, any length, with no per-segment or per-character fee. Current rates are on the Actor's Apify Store page, and the Console shows a live total while a run is in progress.

How do I guarantee an agent never overspends?

Pass maxTotalChargeUsd on the run endpoint — it is a hard per-run ceiling — and cache transcripts by videoId + language so repeats cost nothing.

---

🔍 Companion machine-readable files

File	Purpose
.actor/actor.json	Identity, version, storage and dockerfile wiring
.actor/input_schema.json	Authoritative typed input contract — videoUrl, language enum, translate, proxy editor, deliveryMode enum
.actor/dataset_schema.json	Output field metadata and both Console views (Overview + unwound Transcript) — consumed by the Apify MCP server to describe results to agents
.actor/output_schema.json	Declares where results are stored
my_actor/main.py	Video-ID extraction, caption resolution and parsing, language fallback, MCP delivery
my_actor/connector.py	MCP connector session handling and {placeholder} rendering

If any table on this page disagrees with the schema files, the schema files win.

---

🔗 Related Actors

Other YouTube Actors

This Actor is the single-video specialist: one URL in, one timestamped transcript out. Use these siblings when the job is bigger or different.

Actor	Best for	Why pick it over this one
YouTube Full Channel Transcripts Scraper	Transcripts for every video in a channel — accepts @handle, /channel/ID, /c/name, /user/name, processes up to 50 videos concurrently, and returns video + channel metadata alongside the transcript	You want a whole channel (or its latest N videos) in one run instead of orchestrating one run per video.
YouTube Scraper: Videos, Channels & Comments	Video, channel, and comment metadata from search queries, channel URLs, video URLs, or hashtags — up to 500 videos per input and 1,000 comments per video	You need views, likes, descriptions, channel stats, or comment threads — none of which this transcript Actor returns.
YouTube Shorts Scraper	Every Short from a channel with engagement metrics (views, likes, comments, hashtags) and optional video downloads	You're working through a channel's Shorts catalogue, or you need the video files themselves.

Pick by intent: one video's words → this Actor · a channel's words → Full Channel Transcripts · numbers, comments, and search discovery → YouTube Scraper · a channel's Shorts and their files → Shorts Scraper.

Enrichment & downstream pipeline

Actor	Use it for
Advanced Website Crawling Actor	Crawl the creator's own site or show notes for clean HTML/Markdown/text context alongside the transcript
Google Maps Scraper – Business Data	Enrich businesses or venues mentioned in a transcript into full business records
GOV.UK Business Scraper	Match UK company names heard in a video against Companies House records

Suggested pipeline patterns

✍️ Content repurposing YouTube Transcript Scraper → transcript text into your LLM of choice → blog draft, thread, newsletter → publish via a Notion/Slack connector.

🌍 Localization One run per target language (hi, gu, es, …) → format start/dur/text as SRT downstream → upload as subtitle tracks.

🔎 Video knowledge base YouTube Full Channel Transcripts Scraper for the channel backfill → this Actor for each new upload → store by videoId + language → index searchResult text for semantic search over everything you've published.

📊 Transcript + engagement analysis YouTube Scraper for views, likes, and comments → this Actor for the words of the videos that over-performed → correlate topics against engagement.

🤖 Agent retrieval tool Register the Actor as an MCP tool so an agent can pull a video's transcript on demand and answer questions with timestamps.

Browse all Actors by Techforce Global at scraper.techforce.global.

---

🔐 Compliance & Data Privacy

Trademark & affiliation. This is an independent Actor. It is not affiliated with, endorsed by, or sponsored by YouTube or Google. YouTube™ is a trademark of Google LLC; all trademarks are the property of their respective owners.

Public data only — and no credentials, ever. This Actor reads publicly available caption tracks. It does not log into any account, does not accept or store any Google/YouTube credential, and does not bypass authentication. Private, unlisted and age-gated videos are inaccessible to it by design, and a request for one returns status: "no_transcript" rather than a partial result.

GDPR / CCPA — what you are taking on. The Actor itself collects no personal data about you and requires no account linkage. But a transcript is speech, and speech is about people:

• The speaker is identifiable. A transcript of an interview, vlog or conference talk is that person's words. Under GDPR / UK GDPR and CCPA/CPRA, storing it can constitute processing personal data about them — and about any third party named aloud in the recording.
• Special-category content cannot be filtered in advance. Health, politics, religion and sexuality come up in spoken content, and you cannot know which transcripts contain it until after retrieval. Keep retention short and access controlled.
• Auto-generated transcripts are not verbatim quotes. When isAutoGenerated: true, a machine wrote the text; mis-transcribed names and inverted negations are common. Attributing it to a speaker as their exact words is a misattribution — and potentially a defamation — risk. Check the flag before quoting.
• Minimise. If your question is about topics rather than people, you need the text, not the identity of everyone in it. Aggregate and discard early.

Copyright. Transcripts are the creator's copyrighted expression. Internal research, search indexing and accessibility use sit on far safer ground than republishing transcripts as your own content, or using them as a substitute for watching the video. Using them to train a model is a legally distinct act from analysing them — take advice rather than relying on "it was public".

Terms of Service. YouTube's terms restrict automated access. This Actor does not remove that risk; running it means accepting it. The official Data API is the sanctioned route — but its captions.download endpoint only serves videos you own, which is the gap this Actor exists to fill.

---

🆘 Support & Custom Pipeline Engineering

Whole-channel coverage is already a product — see YouTube Full Channel Transcripts Scraper. Need something beyond it: scheduled multi-channel monitoring, playlist coverage, or a full content pipeline into your CMS?

• Email: bhavin.shah@techforceglobal.com
• Custom Enterprise Integrations: Book a 15-Minute Technical Consultation
• Maintained by: Techforce Global — Specialists in High-Performance Web Scrapers and Agentic Workflows.

---

Made with ❤️ by Techforce Global Specialists in High-Performance Video & Transcript Data Extraction and AI Automation.

---

🏷️ Structured data for search & AI discovery

Two blocks: SoftwareApplication describing the Actor, and FAQPage wrapping the FAQ above so those answers are eligible for rich results. The FAQPage entries are generated from the visible FAQ and match it question-for-question.

{
  "@context": "https://schema.org",
  "@type": "SoftwareApplication",
  "name": "YouTube Transcript Scraper - Captions to Notion, Slack & Jira",
  "alternateName": "YouTube Captions API Alternative",
  "applicationCategory": "DeveloperApplication",
  "operatingSystem": "Cloud (Apify platform)",
  "description": "The YouTube captions API alternative for videos you don't own: extracts the full timestamped transcript of any public YouTube video as structured JSON — every segment with start time, duration and text — in 13 languages with automatic translation fallback, without a Google API key, OAuth, or video ownership. Public caption tracks only: no account login, no stored credentials, and no access to private, unlisted or age-gated videos, with GDPR/CCPA and speaker-identifiability guidance included.",
  "url": "https://apify.com/techforce.global/youtube-transcript-scraper",
  "offers": {
    "@type": "Offer",
    "priceCurrency": "USD",
    "description": "Usage-based: billed for the Apify platform usage a run consumes. One video per run, any length, with no per-segment charge. Current rates are shown on the Actor's Apify Store page."
  }
}

{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "Do I need a YouTube or Google account, API key, or OAuth app?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "No. The Actor reads publicly available caption tracks. You need an Apify account and API token; you need no Google credentials whatsoever."
      }
    },
    {
      "@type": "Question",
      "name": "Is this the official YouTube API?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "No. This is an independent Actor and is not affiliated with, endorsed by, or sponsored by YouTube or Google."
      }
    },
    {
      "@type": "Question",
      "name": "Why can't I just use the official YouTube Data API?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "captions.download only works for videos you own, via OAuth. For any third-party video it is not an option — which is the main reason this Actor exists."
      }
    },
    {
      "@type": "Question",
      "name": "Do I need to configure proxies?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "The residential Apify Proxy is on by default and you should leave it on — YouTube serves datacenter IPs a bot-check wall. Residential groups require a paid Apify plan."
      }
    },
    {
      "@type": "Question",
      "name": "Can I plug it into Claude, Cursor, or a LangChain agent?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Yes — it is a native MCP tool: claude mcp add --transport http apify \"https://mcp.apify.com?tools=techforce.global/youtube-transcript-scraper\""
      }
    },
    {
      "@type": "Question",
      "name": "Can I pass a playlist, channel, or list of videos?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "No — one video per run. For a whole channel, use YouTube Full Channel Transcripts Scraper, which takes a channel URL and processes up to 50 videos concurrently. For a few specific unrelated videos, loop this Actor from your orchestration layer (each video is a separate run, and a separate charge)."
      }
    },
    {
      "@type": "Question",
      "name": "Which languages are supported?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "13: en, hi, gu, pt, es, fr, de, it, ja, ko, zh, ru, ar. Region variants like en-US are not valid inputs."
      }
    },
    {
      "@type": "Question",
      "name": "What happens if my language isn't available?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "With translate: true (default) you get a YouTube auto-translation. With translate: false you get the video's own caption language, fallbackUsed: true, and a message listing what was available. The run does not fail for language reasons."
      }
    },
    {
      "@type": "Question",
      "name": "Does it work on Shorts?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Yes — Shorts, embed, /live/, youtu.be, mobile and youtube-nocookie hosts, and bare 11-character IDs all resolve."
      }
    },
    {
      "@type": "Question",
      "name": "Why is the whole transcript in one dataset item?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "One run handles one video, so the transcript is a nested searchResult array on a single item. Use the Console's Transcript view (or unwind=searchResult on the API) for a per-segment table."
      }
    },
    {
      "@type": "Question",
      "name": "Why are start and dur strings?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "For compatibility with the widely-used searchResult transcript format. Cast them to float before doing arithmetic."
      }
    },
    {
      "@type": "Question",
      "name": "Does it summarize or translate with an LLM?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "No. It returns YouTube's caption text — raw, timestamped, and unmodified apart from HTML unescaping. Summarize downstream."
      }
    },
    {
      "@type": "Question",
      "name": "Can it transcribe a video that has no captions?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "No. There is no speech recognition in this Actor. Use an ASR service for caption-less videos. The run itself still succeeds and tells you so, with status: \"no_transcript\" and errorCode: \"NO_CAPTIONS\"."
      }
    },
    {
      "@type": "Question",
      "name": "Why did my run succeed but return no transcript?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Because the video, not the Actor, was the problem: it may be private, deleted, region-blocked, caption-less, or YouTube may have challenged the request. Rather than crashing, the Actor writes one dataset item with status: \"no_transcript\", a machine-readable errorCode, and a plain-English message — so scheduled runs and agent pipelines get a readable reason instead of an alert storm from failed runs. Read item[\"status\"] first, always."
      }
    },
    {
      "@type": "Question",
      "name": "Which outcomes still fail the run?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "None of them. A missing videoUrl (MISSING_VIDEO_URL), an unusable one (BAD_VIDEO_URL) and every fact about the video's own state all end as a SUCCEEDED run whose single item carries an errorCode and a validation message — so API and MCP callers get a readable reason instead of a stack trace. A FAILED run now means a platform-level problem: an abort, a timeout, or an exhausted spend cap."
      }
    },
    {
      "@type": "Question",
      "name": "What is the MCP connector for?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "It pushes the transcript straight into a tool you have already authorized on your Apify account — Notion, Slack, Jira, GitHub, Linear, Airtable, or any MCP-compatible connector — with no glue code. Credentials stay private: delivery runs through the Apify MCP Proxy, so the Actor never sees your connector tokens."
      }
    },
    {
      "@type": "Question",
      "name": "Which delivery mode should I use?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "summary for short videos, chunked for long transcripts into services with request caps (Notion), perSegment for per-line records (logs, row-per-segment sheets), none for dataset-only runs."
      }
    },
    {
      "@type": "Question",
      "name": "Does delivery work when running locally?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "No. Connectors resolve through the Apify MCP Proxy, which only exists on the platform. Locally you'll see APIFY_MCP_PROXY_URL is not set and the dataset item is saved as normal."
      }
    },
    {
      "@type": "Question",
      "name": "Does it deliver anything when there is no transcript?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "No. A no_transcript outcome skips delivery entirely, so your Slack channel or Notion database never receives an empty transcript. The reason is in the dataset item and the run's status message."
      }
    },
    {
      "@type": "Question",
      "name": "What does it cost?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Usage-based: you pay for the platform usage a run consumes. One video per run, any length, with no per-segment or per-character fee. Current rates are on the Actor's Apify Store page, and the Console shows a live total while a run is in progress."
      }
    },
    {
      "@type": "Question",
      "name": "How do I guarantee an agent never overspends?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Pass maxTotalChargeUsd on the run endpoint — it is a hard per-run ceiling — and cache transcripts by videoId + language so repeats cost nothing."
      }
    }
  ]
}