# yTranscript > yTranscript (ytranscript.com) gets the transcript of any YouTube video. It offers a free web app, a REST API for developers (transcripts and video metadata), an MCP server for AI assistants (Claude, Cursor, VS Code), bulk export of whole channels and playlists, and an Apify actor. It works from cloud servers without proxies, returns the language actually spoken by default, and transcribes videos that have no captions using AI speech-to-text (Whisper). Key facts: - Problem it solves: YouTube blocks most cloud and datacenter IPs (AWS, GCP, Vercel, Lambda, n8n Cloud) from fetching transcripts, so open-source libraries such as youtube-transcript-api often fail in production with "RequestBlocked", "IpBlocked" or "Sign in to confirm you're not a bot". yTranscript fetches on your behalf, so your servers never contact YouTube and you don't need to buy proxies. - Videos without captions are transcribed with AI automatically, through the same endpoint. - Failed requests are never charged. - API pricing is unit-based. A captioned video costs 1 unit. AI transcription of a video with no captions costs 1 unit per minute of audio, with a minimum of 5 units (a 2-hour video, the longest accepted, costs 120). Video metadata costs 1 unit, or nothing extra when requested together with a transcript. The price a customer signs up at stays theirs while they remain subscribed. Plans: Free $0 (50 units/month, no card), Starter $9/month (1,000 units), Developer $29/month (10,000 units), Scale $79/month (50,000 units). Pay-as-you-go packs add 1,000 units for $10, 5,000 for $40 or 20,000 for $120 on any plan. - The web app is free to use with a daily limit. Pro costs $4.99/month, removes the limit, and adds AI transcription for videos without captions and AI summaries. ## Developer API - [YouTube Transcript API](https://ytranscript.com/developers?utm_source=llms.txt): REST API docs for transcripts and video metadata, pricing and free API key signup - [How to extract YouTube transcripts at scale](https://ytranscript.com/blog/youtube-transcript-api?utm_source=llms.txt): why transcript libraries get blocked on cloud servers and how to avoid it Quick reference: ``` GET https://ytranscript.com/api/v1/transcript?videoId=dQw4w9WgXcQ Authorization: Bearer yk_live_YOUR_KEY ``` Parameters: `videoId` (11-character ID) or `url` (any YouTube URL), plus optional `lang` (caption language code such as `en` or `es`; defaults to the video's original language) optional `metadata=true` (adds a `metadata` object with the video metadata fields below at no extra cost; `null` if it can't be read) and optional `max_units` (the most the request may cost: if the video has no captions and AI transcription would cost more, the API answers 422 `max_units_exceeded` with `estimated_units` before transcribing and charges nothing; `max_units=1` means captions only). Response (JSON): `video_id`, `lang`, `whisper` (true if AI-transcribed), `cached`, `units_charged` (what the request cost; a cached repeat within 24 hours costs 1), `audio_seconds` (on live AI transcriptions: the audio length the charge is based on), and `segments`, an array of `{ text, offset, duration }` with times in milliseconds. Errors are JSON: `{ "error": { "code", "message", ...extra fields } }`. Branch on `code`: 400 `bad_request` (no `videoId` or `url`, not a YouTube video, or a bad `max_units`), 401 `unauthorized` (missing, invalid or revoked key), 402 `quota_exceeded` (monthly units used up and no top-up units; extra fields `quota`, `used`, `top_up_units`), 404 `no_transcript` (the video has no transcript we can get), 422 `max_units_exceeded` (AI transcription would cost more than the request's `max_units`; extra fields `max_units`, `estimated_units`, which is null when the length can't be read), 429 `rate_limited` (too many requests this minute; extra field `retry_after_seconds`), 500 `fetch_failed` (unexpected failure; retry), 503 `temporarily_unavailable` (YouTube is temporarily blocking us: retry after the `Retry-After` header or `retry_after_seconds`; the video is not missing). Failed requests are not charged. Caption transcripts usually return in 2 to 3 seconds. AI transcription (`whisper: true`) downloads the audio first and can take several minutes on long videos, so use a client timeout of at least 300 seconds. Every response includes the `X-Quota-Limit`, `X-Quota-Used` and `X-RateLimit-Remaining` headers, plus `X-TopUp-Units-Remaining` once the monthly units are used up and requests draw on top-up units. Video metadata: ``` GET https://ytranscript.com/api/v1/video?videoId=dQw4w9WgXcQ Authorization: Bearer yk_live_YOUR_KEY ``` Response (JSON): `video_id`, `title`, `description`, `duration_seconds`, `view_count`, `like_count`, `published_at`, `channel` (`id`, `name`, `handle`, `url`, `subscriber_count`), `tags`, `category`, `chapters` (array of `{ title, start_seconds }`), `thumbnails` (array of `{ url, width, height }`), `is_live`, `cached` and `units_charged`. Costs 1 unit. `duration_seconds`, `tags` and `category` can be `null`, and `published_at` can be a date without a time, when YouTube limits what it shows our servers. Errors are the same as for transcripts, except that 404 is `not_found` (the video is private, removed or does not exist). ## MCP server (Claude, Cursor, VS Code) - [ytranscript-mcp on GitHub](https://github.com/nnnoel/ytranscript-mcp): setup for every MCP client; two tools: `get_transcript` (plain or timestamped output, an optional length cap, and optional video details at no extra cost) and `get_video_info` (title, description, channel, publish date, duration, views, likes, tags, category and chapters) - [ytranscript-mcp on npm](https://www.npmjs.com/package/ytranscript-mcp): run with `npx -y ytranscript-mcp` and the `YTRANSCRIPT_API_KEY` environment variable Claude Code: `claude mcp add ytranscript -e YTRANSCRIPT_API_KEY=yk_live_YOUR_KEY -- npx -y ytranscript-mcp` ## Libraries and integrations - [ytranscript-api on npm](https://www.npmjs.com/package/ytranscript-api): JavaScript/TypeScript client for the API: `yt.transcript()` and `yt.video()` - [YouTube Transcript Scraper on Apify](https://apify.com/ytranscript/youtube-transcript): bulk transcripts with pay-per-result pricing, optional video details (title, channel, views, chapters) in each row, no yTranscript account needed ## Web app - [yTranscript](https://ytranscript.com/?utm_source=llms.txt): paste a YouTube link and copy or download the transcript, with timestamps, language selection and AI summaries - [Bulk transcript download](https://ytranscript.com/bulk?utm_source=llms.txt): export every transcript from a channel or playlist as a zip (TXT, SRT or JSON) with a manifest.csv ## Guides - [How to download a YouTube transcript](https://ytranscript.com/blog/how-to-download-youtube-transcript?utm_source=llms.txt) - [Fix YouTube transcript not loading](https://ytranscript.com/blog/fix-youtube-transcript-not-loading?utm_source=llms.txt) - [Download YouTube captions](https://ytranscript.com/blog/youtube-captions-download?utm_source=llms.txt) - [Convert a YouTube video to text](https://ytranscript.com/blog/convert-youtube-video-to-text?utm_source=llms.txt) - [YouTube video summarizer](https://ytranscript.com/blog/youtube-video-summarizer?utm_source=llms.txt) ## Optional - [Privacy policy](https://ytranscript.com/privacy?utm_source=llms.txt)