YouTube Transcript API: Scrape Video Transcripts, No API Key

Fetch the transcript of any public YouTube video, Short or embed as clean text, timed JSON segments, SRT or WebVTT. Manual captions are used before auto-generated ones, and every video gets either a transcript or a clear error code instead of failing the run. Pay per use: $4 per 1,000 transcripts, failed videos are free.

Use it to download YouTube transcripts in bulk, feed videos to ChatGPT, Claude or a RAG pipeline, or build a searchable transcript archive without a YouTube API key.

This Actor runs through Apify residential proxy on the Apify platform, because YouTube blocks most datacenter and cloud IPs. The proxy is enabled by default in the input and its cost is included in the per-transcript price.

Use it from an AI agent (MCP)

Call this Actor from any MCP client through the Apify MCP server. A hosted MCP server and HTTP API for the same data is being prepared at api.openkrill.app (planned tool get_transcript, overview at tools.openkrill.app); the YouTube tool there is coming soon.

What it does

• Accepts video IDs and every common link form: watch?v=, youtu.be/, /shorts/, /embed/, /live/, m. and music. hosts.
• Picks the best caption track by your language priority list, using human-made captions before auto-generated ones for each language; en also matches regional tracks such as en-US or en-GB, with an exact match winning.
• When none of your languages exist, returns the transcript in the video's original spoken language instead of failing (languageFallback: true marks it); set fallbackToAnyLanguage to false to get a NoTranscriptFound error instead.
• Optionally returns YouTube's own machine translation into a target language.
• Returns the video title and channel name with every transcript.
• Never fails the whole run because of one video: each video gets either a transcript or an error object with a machine-readable code and a retryable flag.
• Skips duplicate videos in the same run, so you are never charged twice for one transcript.
• Retries network and server errors with exponential backoff, and retries blocked requests from fresh proxy IPs when a proxy is configured.

Input

Field	Type	Default	Description
videos	string[]	required	Video URLs or IDs, up to 1000 per run.
languages	string[]	["en"]	Language codes in priority order, e.g. ["en", "en-GB", "de"]. A code also matches its regional variants.
fallbackToAnyLanguage	boolean	true	When none of languages exist, return the original-language (or first available) track instead of an error.
translateTo	string	none	Language code to translate into when no track in that language exists.
outputFormat	segments | text | srt | vtt	segments	Shape of the transcript in each result.
maxConcurrency	integer 1-10	3	Videos fetched in parallel.
proxyConfiguration	proxy	Apify residential	Leave the default on the Apify platform; see Limits.

Example:

{
  "videos": [
    "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
    "https://youtu.be/jNQXAC9IVRw",
    "https://www.youtube.com/shorts/xxxxxxxxxxx"
  ],
  "languages": ["en"],
  "outputFormat": "segments"
}

Output

One dataset item per video. A successful item (segments shortened):

{
  "status": "ok",
  "input": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
  "videoId": "dQw4w9WgXcQ",
  "title": "Rick Astley - Never Gonna Give You Up (Official Video) (4K Remaster)",
  "channelName": "Rick Astley",
  "languageCode": "en",
  "language": "English",
  "isGenerated": false,
  "languageFallback": false,
  "translatedFrom": null,
  "format": "segments",
  "text": "[♪♪♪] ♪ We're no strangers to love ♪ ♪ You know the rules and so do I ♪ ...",
  "segments": [
    { "text": "[♪♪♪]", "start": 1.36, "duration": 1.68 },
    { "text": "♪ We're no strangers to love ♪", "start": 18.64, "duration": 3.24 }
  ],
  "segmentCount": 61,
  "fetchedAt": "2026-09-30T08:00:00.000Z"
}

With outputFormat srt or vtt, segments is replaced by subtitles holding the file content; with text, only text is returned. text is always present.

A failed item:

{
  "status": "error",
  "input": "https://www.youtube.com/shorts/xxxxxxxxxxx",
  "videoId": "xxxxxxxxxxx",
  "error": {
    "code": "VideoUnavailable",
    "message": "The video is no longer available: xxxxxxxxxxx",
    "retryable": false
  },
  "fetchedAt": "2026-09-30T08:00:00.000Z"
}

Error codes

Code	Meaning	Retryable
InvalidVideoId	The input is not a YouTube video ID or link.	no
VideoUnavailable	The video does not exist or was removed.	no
VideoUnplayable	YouTube will not play it (private, processing, region-locked); the message has YouTube's reason.	no
AgeRestricted	Age-gated; transcripts need a signed-in account, which this Actor does not use.	no
TranscriptsDisabled	The uploader turned captions off.	no
NoTranscriptFound	None of your languages exist and fallbackToAnyLanguage is false; error.available lists the ones that do.	no
NotTranslatable / TranslationLanguageNotAvailable	YouTube does not offer that translation; error.available lists options.	no
RequestBlocked / IpBlocked	YouTube blocked the request as automated traffic.	yes
PoTokenRequired	YouTube demanded a proof-of-origin token for this track.	no
FailedToCreateConsentCookie	The EU cookie consent page could not be passed.	no
YouTubeRequestFailed	Network or YouTube server error after retries.	yes
YouTubeDataUnparsable	YouTube changed its response format; we monitor for this and fix it.	no

Pricing

Pay per event: $4 per 1,000 transcripts ($0.004 per transcript event, charged only for items with status: "ok"). Failed videos, invalid inputs and duplicates are free. Residential proxy traffic is covered by that price; you are not billed for it separately. On the Apify free plan a run processes the first 10 videos and skips the rest (the run log and status message say so); any paid Apify plan runs up to 1000 videos per run.

FAQ

Do I need a YouTube API key? No.

Does it work for videos without captions? No. Only captions YouTube already has are returned; there is no speech-to-text.

Can I get channel or playlist transcripts? Not directly: pass the video URLs or IDs. Up to 1000 videos per run.

Which formats are supported? segments (timed JSON plus text), text, srt and vtt.

Limits

• YouTube blocks most datacenter and cloud IPs, so the Actor is meant to run on the Apify platform with the default residential proxy. Turning the proxy off (or running it elsewhere) typically produces RequestBlocked or IpBlocked on anything beyond a few videos.
• Translation uses YouTube's own machine translation. YouTube currently rate-limits translated caption downloads for anonymous clients, so translateTo often returns RequestBlocked; requesting the original language is reliable.
• Age-restricted, private and members-only videos are not supported, because they require signing in.
• On the Apify free plan, runs are capped at 10 videos; upgrade to any paid plan for up to 1000 per run.
• Only captions YouTube already has are returned; there is no speech-to-text for videos without captions.

Credits

The caption discovery approach is a TypeScript port of youtube-transcript-api by Jonas Depoix (MIT License). See NOTICE for the license text.