YouTube Video Transcript and Subtitles Scraper API
Pricing
from $0.80 / 1,000 transcript extracteds
YouTube Video Transcript and Subtitles Scraper API
YouTube video transcript scraper and subtitles API. Get transcripts, captions, timestamps and video metadata in bulk from videos, Shorts, playlists and channels. Pay per video, no API key needed. Callable from any MCP client such as Claude, Cursor or ChatGPT. Built for RAG and LLM datasets.
Pricing
from $0.80 / 1,000 transcript extracteds
Rating
0.0
(0)
Developer
George Kioko
Maintained by CommunityActor stats
1
Bookmarked
430
Total users
12
Monthly active users
a day ago
Last modified
Categories
Features
Share
YouTube Transcript and Subtitles Scraper API is an Apify Actor that extracts transcripts, subtitles, timed caption segments and video metadata from YouTube videos, playlists, channels and Shorts, and returns them as structured JSON. It reads YouTube caption tracks directly, so you need no API key and no YouTube Data API quota. It works as a bulk YouTube transcript API, a real time HTTP transcript endpoint and a YouTube subtitles scraper for RAG datasets, LLM training corpora, content repurposing and video SEO analysis.
No API key. No YouTube Data API quota. No monthly subscription. You pay per extracted transcript. Videos with no captions at all never trigger a per transcript charge.
Pricing
| Event | Price | You pay when |
|---|---|---|
| Actor start | $0.0001 | Once per run, or once per HTTP request in Standby |
| Transcript extracted | $0.001 | Per video transcript successfully extracted |
$1 per 1,000 transcripts. The start fee is a hundredth of a cent. A 500 video channel costs about $0.50. Pay only for transcripts you actually receive.
Why this Actor?
- No setup. No Google Cloud project, no OAuth, no API quota limits.
- Bulk by default. Up to 5,000 videos per run. Mix video URLs, playlist URLs, channel URLs, Shorts and raw IDs in one input.
- Real time HTTP API. One GET request returns the transcript in the reply, see Standby below.
- Auto generated captions. Works even when the creator never uploaded manual subtitles.
- Language fallback. Ask for
en, and if it does not exist you still get the best available track instead of an empty row. - LLM ready output. Full text, timestamped segments or both, plus title, channel, views and duration on every row.
How it works
flowchart LRA[Video, playlist,<br/>channel, Shorts URL] --> B[Normalize URL<br/>and discover videos]B --> C[Fetch transcript tracks<br/>manual or auto captions]C --> D{Preferred language<br/>available?}D -- yes --> E[Extract selected<br/>language transcript]D -- no --> F[Fallback to best<br/>available captions]E --> G[Format output<br/>full text, segments, or both]F --> GG --> H[Attach metadata<br/>title, channel, views,<br/>duration, thumbnail]H --> I[Dataset export or<br/>HTTP reply]
What data does it extract?
- Full transcript text in
transcriptText, withwordCountandcharCount - Timed segments in
segments, each with start, duration and a formatted start time - Video metadata such as title, channel name and ID, view count, publish date, length, category, keywords and thumbnail
- Languages with the selected track and every track YouTube offers for the video
- Video URL and ID for the source video
What is NOT returned
- Video or audio files. This extracts text, not media.
- Comments or community posts.
- Machine translation. You get the caption tracks YouTube actually serves, in their original languages.
- Burned in subtitles. Text rendered into the video image has no caption track to read.
Quick start
- Open the YouTube Transcript Scraper on Apify Store and click Try for free.
- Paste video URLs, playlist URLs, channel URLs or raw video IDs into URLs. Formats can be mixed.
- Pick a Language (default
en, falls back automatically), an Output Format (full-text,segmentsorboth), and toggle timestamps and metadata. - Click Start, then export the dataset as JSON, CSV or Excel, or pull it through the API.
Real time HTTP API (Standby)
The actor also runs as a warm web server. Send one GET request with a video and the transcript comes back in the reply. There is no run to start and no dataset to fetch, which suits AI agents, chat apps and any tool that needs one transcript at a time.
curl "https://george-the-developer--youtube-transcript-scraper.apify.actor/?url=https://www.youtube.com/watch?v=jNQXAC9IVRw&outputFormat=full-text" \-H "Authorization: Bearer $APIFY_TOKEN"
{ "count": 1, "rows": [ { "videoId": "jNQXAC9IVRw", "title": "Me at the zoo", "hasTranscript": true, "wordCount": 39, "transcriptText": "All right, so here we are, in front of the elephants ..." } ] }
Pass url, videoUrl or videoId once or many times, up to 10 videos per request. language, outputFormat, includeTimestamps and includeMetadata work as query parameters too. A POST with the same JSON as a normal run also works. The price is the same as a normal run. Playlists and channels go through a normal run.
Input parameters
| Parameter | Type | Required | Description |
|---|---|---|---|
urls | Array | Yes | YouTube video URLs, playlist URLs, channel URLs or video IDs |
language | String | No | Preferred transcript language code (default en) |
outputFormat | Enum | No | full-text, segments or both (default both) |
includeTimestamps | Boolean | No | Include start times and durations for each segment (default true) |
maxVideos | Integer | No | Maximum videos to process, up to 5,000 (default 50) |
includeMetadata | Boolean | No | Include video title, channel, views and more (default true) |
maxConcurrency | Integer | No | Concurrent requests, 1 to 20 (default 5) |
proxyConfiguration | Object | No | Apify Proxy country routing only. The actor always uses BUYPROXIES94952, custom proxy URLs and other groups are ignored |
A run with an empty urls list stops at once with a clear message and costs nothing.
Output example
A real row from 2026-11-18, with the lists shortened.
{"videoId": "jNQXAC9IVRw","videoUrl": "https://www.youtube.com/watch?v=jNQXAC9IVRw","language": "en","languageName": "English","isAutoGenerated": false,"availableLanguages": [{ "code": "en", "name": "English", "isAutoGenerated": false },{ "code": "de", "name": "German", "isAutoGenerated": false }],"hasTranscript": true,"title": "Me at the zoo","channelName": "jawed","channelId": "UC4QobU6STFB0P71PMvOGN5A","lengthSeconds": 19,"viewCount": 448788606,"publishDate": "2005-04-23T20:31:52-07:00","category": "Film & Animation","keywords": ["me at the zoo", "jawed karim", "first youtube video"],"transcriptText": "All right, so here we are, in front of the elephants the cool thing about these guys is that they have really... really really long trunks and that's cool (baaaaaaaaaaahhh!!) and that's pretty much all there is to say","charCount": 217,"wordCount": 39,"segments": [{ "text": "All right, so here we are, in front of the elephants", "start": 1.2, "duration": 2.16, "startFormatted": "0:01" },{ "text": "the cool thing about these guys is that they have really...", "start": 5.318, "duration": 2.656, "startFormatted": "0:05" }],"segmentCount": 6,"scrapedAt": "2026-11-18T13:27:57.754Z"}
Use from Claude, ChatGPT and any MCP agent
Add this actor to any MCP client (Claude, Cursor, ChatGPT) through the Apify MCP server.
https://mcp.apify.com/?tools=george.the.developer/youtube-transcript-scraper
Or call it from code.
import { ApifyClient } from 'apify-client';const client = new ApifyClient({ token: 'YOUR_TOKEN' });const run = await client.actor('george.the.developer/youtube-transcript-scraper').call({urls: ['https://www.youtube.com/watch?v=dQw4w9WgXcQ'],outputFormat: 'both',});const { items } = await client.dataset(run.defaultDatasetId).listItems();// items[0].transcriptText, items[0].segments
curl -X POST "https://api.apify.com/v2/acts/george.the.developer~youtube-transcript-scraper/run-sync-get-dataset-items?token=YOUR_TOKEN" \-H "Content-Type: application/json" \-d '{"urls": ["https://www.youtube.com/watch?v=dQw4w9WgXcQ"]}'
Integrations
- n8n. Add the Apify node, pick Run Actor and Get Dataset Items, set the actor to
george.the.developer/youtube-transcript-scraperand maptranscriptTextinto your workflow. Or call the Standby URL with an HTTP Request node. - Make and Zapier. Use the Apify app Run Actor module, then Get Dataset Items.
- Anything else. Every run's dataset has a stable URL,
https://api.apify.com/v2/datasets/<DATASET_ID>/items?token=<TOKEN>&format=json&clean=1.
AI and content pipeline
flowchart TDYT[YouTube videos] --> TR[Transcript extraction]TR --> CLEAN[Clean text + timestamps]CLEAN --> RAG[RAG / vector database]CLEAN --> SEO[Video SEO analysis]CLEAN --> REP[Blog posts, newsletters,<br/>show notes, clips]RAG --> APP[Chatbot or research assistant]SEO --> PLAN[Keyword and topic gaps]REP --> CMS[CMS / social scheduler]
Use cases
- RAG and LLM fine tuning. Turn hundreds of domain videos into a knowledge base or training corpus.
- Content repurposing at scale. One video becomes blog posts, threads, newsletters and show notes.
- Video SEO and competitor analysis. Mine the spoken content of top ranking videos for keyword and topic gaps.
- Podcast and webinar archives. Long recordings become timestamped searchable text.
- Creator research and clipping. Find recurring hooks and send timestamped sections to editors or AI agents.
FAQ
Do I need a YouTube Data API key?
No. This scraper works without any API key or Google account. It reads caption tracks directly, so YouTube API quotas never apply.
How much does it cost to transcribe 1,000 videos?
$1. Pricing is $0.001 per extracted transcript plus a $0.0001 start fee. Videos without captions never trigger a per transcript charge.
Does it work with auto generated captions?
Yes. Both manually uploaded subtitles and YouTube's auto generated captions are supported, in any language YouTube serves.
Can I scrape entire playlists or channels?
Yes. Pass a playlist or channel URL in a normal run and the scraper discovers and processes every video, up to your maxVideos limit (max 5,000 per run).
Can I get a transcript in real time without starting a run?
Yes. Use the Standby URL shown above. One GET request returns up to 10 videos in the reply at the same price as a normal run.
What languages are supported?
Every language YouTube provides a caption track for. Set your preferred language code. If it is not available the scraper falls back to the best available track instead of returning nothing.
How do I handle geo restricted videos?
Set proxyConfiguration.countryCode to route through Apify Proxy in the right country. The actor always uses the BUYPROXIES94952 proxy group for reliability.
Can I use the output for LLM training data?
Yes. The full-text format suits direct ingestion into fine tuning and RAG pipelines. Process up to 5,000 videos per run to build large corpora.
Which output formats can I choose?
Set outputFormat to full-text for the transcript as one string, segments for the list of timed segments, or both, which is the default. Timestamps are included on each segment unless you turn includeTimestamps off. Video metadata such as title, channel, view count and length comes with includeMetadata, which is also on by default.
Can an AI agent call this actor?
Yes. Add it to any MCP client, such as Claude, Cursor or ChatGPT, through the Apify MCP server at https://mcp.apify.com/?tools=george.the.developer/youtube-transcript-scraper. Agents that make plain HTTP calls can use the Standby URL, and it also runs from the Apify API with curl or from JavaScript with apify-client.