# YouTube Video Transcript

> YouTube Video Transcript provides transcripts, video metadata, search, and comments through a website, public REST API, and MCP server. Collect transcripts or standalone comments from videos, playlists, and channels in background jobs. One balance: 1 credit per successful transcript and 1 per nonempty comments page, usually about 20 top-level comments. Free accounts receive 30 credits every 30 days after Google sign-in.

Canonical URL: https://youtubevideotranscript.io

Built for researchers, content creators, journalists, and students who need searchable transcripts and comment datasets. Search, video details, channel and playlist listings, and reads of saved job results cost zero credits. Free requests remain rate-limited.

## Core pages

- [Homepage](https://youtubevideotranscript.io): transcripts, metadata, comments, search, and background collections
- [Bulk transcripts](https://youtubevideotranscript.io/bulk): channels, playlists, or a custom selection, with optional Top comment samples
- [Comments collection](https://youtubevideotranscript.io/comments): standalone comments, Top or Newest, an amount per video or all available, optional replies, and a credit limit
- [MCP server](https://youtubevideotranscript.io/mcp): fourteen tools for AI chats
- [Pricing](https://youtubevideotranscript.io/pricing): Free (30 credits every 30 days), Starter ($9/mo for 1,000), Pro ($19/mo for 5,000), Business ($49/mo for 20,000)
- [Free plan](https://youtubevideotranscript.io/free): recurring shared credits after Google sign-in, no credit card
- [YouTube channel transcript](https://youtubevideotranscript.io/youtube-channel-transcript): collect long-form transcripts within the job cap and available credits
- [YouTube playlist transcript](https://youtubevideotranscript.io/youtube-playlist-transcript): collect transcripts from a playlist
- [Transcript API](https://youtubevideotranscript.io/transcript-api): programmatic access with an API key from the account dashboard

## Documentation

- [Docs overview](https://youtubevideotranscript.io/docs): complete public API endpoint index and links to website and MCP guides
- [Using the website](https://youtubevideotranscript.io/docs/website): Workbench, search, transcript collections, standalone comments jobs, and History
- [Authentication](https://youtubevideotranscript.io/docs/authentication): Bearer API keys and rotation
- [Credits and refunds](https://youtubevideotranscript.io/docs/quotas): shared balance, successful fetches, reservations, and refunds
- [Rate limits](https://youtubevideotranscript.io/docs/rate-limits): per-account API/MCP request budgets
- [Idempotency](https://youtubevideotranscript.io/docs/idempotency): safe REST job retries; sync transcript and comments fetches can charge again
- [Search](https://youtubevideotranscript.io/docs/search): YouTube-wide and in-channel search, filters, and pagination
- [Video info](https://youtubevideotranscript.io/docs/video-info): caption tracks, numeric views, and optional dates, likes, and category
- [Resolve](https://youtubevideotranscript.io/docs/resolve): channel and playlist video listings
- [Channel utilities](https://youtubevideotranscript.io/docs/channel): public playlists and exact long-form counts
- [Transcripts](https://youtubevideotranscript.io/docs/transcript): sync retrieval, language choice, metadata, and formats
- [Comments](https://youtubevideotranscript.io/docs/comments): paid nonempty pages, continuation tokens, and reply threads
- [Jobs](https://youtubevideotranscript.io/docs/jobs): submit transcripts, list history, read saved results, cancel, and export either job kind
- [Export formats](https://youtubevideotranscript.io/docs/formats): transcript files, comment samples, and standalone JSON/CSV archives
- [Usage](https://youtubevideotranscript.io/docs/usage): shared balance with legacy transcripts_* field names
- [Errors](https://youtubevideotranscript.io/docs/errors): REST error types and request IDs
- [MCP reference](https://youtubevideotranscript.io/docs/mcp): all tools, parameters, billing, and differences from REST
- [Claude setup](https://youtubevideotranscript.io/docs/mcp/claude)
- [ChatGPT setup](https://youtubevideotranscript.io/docs/mcp/chatgpt)
- [Grok setup](https://youtubevideotranscript.io/docs/mcp/grok)
- [Changelog](https://youtubevideotranscript.io/docs/changelog)

## Blog

- [Blog index](https://youtubevideotranscript.io/blog)
- [Best YouTube transcript tools for AI and LLM datasets (2026)](https://youtubevideotranscript.io/blog/best-youtube-transcript-tools-ai-llm-datasets-2026)
- [How to download every transcript from a YouTube channel (2026)](https://youtubevideotranscript.io/blog/how-to-download-all-youtube-channel-transcripts-2026)
- [Best YouTube transcript APIs compared (2026)](https://youtubevideotranscript.io/blog/best-youtube-transcript-apis-2026)
- [Best YouTube transcript downloaders compared (2026)](https://youtubevideotranscript.io/blog/best-youtube-transcript-downloaders-2026)
- [YouTube transcript for AI training](https://youtubevideotranscript.io/blog/youtube-transcript-for-ai-training)
- [How to download YouTube transcripts](https://youtubevideotranscript.io/blog/how-to-download-youtube-transcripts)
- [How to download YouTube subtitles](https://youtubevideotranscript.io/blog/how-to-download-youtube-subtitles)
- [How to download YouTube captions](https://youtubevideotranscript.io/blog/how-to-download-youtube-captions)
- [Free YouTube transcript generator](https://youtubevideotranscript.io/blog/free-youtube-transcript-generator)
- [Extract text from a YouTube video](https://youtubevideotranscript.io/blog/extract-text-from-youtube-video)

## Use cases

- [For AI and LLM researchers](https://youtubevideotranscript.io/for/researchers): bulk JSON with timestamps for fine-tuning, eval, and RAG corpora
- [For content creators](https://youtubevideotranscript.io/for/creators): repurpose back catalogs into blog posts, newsletters, and threads
- [For journalists](https://youtubevideotranscript.io/for/journalists): searchable transcript archives of press conferences and public talks
- [For students and educators](https://youtubevideotranscript.io/for/students): turn lecture playlists into searchable study notes

## Comparisons

- [vs youtube-transcript.io](https://youtubevideotranscript.io/vs/youtube-transcript-io): subscription versus tokens, side by side
- [vs NoteGPT](https://youtubevideotranscript.io/vs/notegpt): bulk transcripts versus AI summaries
- [vs downloadyoutubetranscripts.com](https://youtubevideotranscript.io/vs/downloadyoutubetranscripts): subscription versus one-time credit packs
- [vs Tactiq](https://youtubevideotranscript.io/vs/tactiq): bulk YouTube versus per-seat meeting transcription
- [Best youtube-transcript.io alternatives (2026)](https://youtubevideotranscript.io/blog/youtube-transcript-io-alternatives-2026): bulk, API, and DIY alternatives by use case
- [Best NoteGPT alternatives for YouTube transcripts (2026)](https://youtubevideotranscript.io/blog/notegpt-alternatives-youtube-transcripts-2026): bulk transcript extraction alternatives to an AI note-taker

## Key facts

- Pricing model: monthly subscription (not credit packs)
- Free tier: 30 shared credits every 30 days, recurring, requires Google sign-in
- Paid tiers: $9/mo (1,000 credits), $19/mo (5,000), $49/mo (20,000)
- Successful transcript: 1 credit. Nonempty comments or reply page: 1 credit. Empty pages and failed fetches: no charge. Repeated successful fetches charge again.
- Transcript export formats: TXT, SRT, JSON, CSV, Markdown in the website; DOCX also available through the API and MCP-created transcript jobs
- Website transcript collections: available on all plans within the balance; whole-channel jobs cap at 5,000 long-form videos and exclude Shorts
- API/MCP transcript bulk creation: paid plans only; per-job caps of 500 / 2,000 / 10,000 for Starter / Pro / Business
- REST POST /jobs enumerates before charging, rejects oversized or unaffordable jobs, and requires Idempotency-Key. MCP channel jobs reserve first and may clip to the plan cap or available balance.
- Optional bulk comments: one Top page per successful transcript, no replies; reserve 2 credits per video and refund unused reservations
- Standalone comments collection: started in the website, up to 5,000 source videos, one active comments job per account; Top/Newest, amount or all available, optional replies, hard credit budget
- Standalone comments exports: ZIP with manifest and per-video JSON plus CSV. Budget-limited completed jobs can contain partial coverage.
- API and MCP can fetch individual comments/reply pages and read existing jobs; they do not create standalone comments collection jobs
- API access: available with a dashboard-generated key, including on Free. Base URL: https://api.youtubevideotranscript.io/api/v1
- MCP endpoint: https://api.youtubevideotranscript.io/mcp (Streamable HTTP; OAuth or Bearer API key)
- Caption sources: manually uploaded captions and YouTube auto-generated captions
- Caption availability: get_video_info or GET /video/:videoId reports available tracks without fetching a transcript; availability does not guarantee a later fetch succeeds
- Single-video REST transcript metadata does not include views, dates, likes, or comments; use the dedicated details and comments endpoints. MCP get_transcript can optionally include one comments page.
- Saved-job reads and repeated downloads cost no additional credits
- History lists jobs created in the last 100 days; exports expire 100 days after a job becomes terminal