![\"TranscriptOut\"](\"public/brand/logo-512.png\"){kind=logo}
\n \n\n \n \n \n
\n Let your AI agent read YouTube.
\n Transcripts in five formats, native video and channel search, channel and playlist listings,\n and background jobs that deliver up to 4,000 transcripts at a time. 14 tools behind one hosted\n endpoint. Nothing runs on your machine.\n
\n ![\"Install](\"https://img.shields.io/badge/Cursor-Install_MCP-000000?style=for-the-badge&logo=cursor&logoColor=white\"/){kind=icon}
![\"Install](\"https://img.shields.io/badge/VS_Code-Install_MCP-0098FF?style=for-the-badge&logo=visualstudiocode&logoColor=white\"/){kind=icon}
\n ![\"Website\"/](\"https://img.shields.io/badge/Website-transcriptout.com-FF3B00?style=for-the-badge\"){kind=icon}
![\"Docs\"/](\"https://img.shields.io/badge/Docs-API_Reference-06B6D4?style=for-the-badge&logo=readthedocs&logoColor=white\"){kind=icon}
![\"MIT](\"https://img.shields.io/badge/License-MIT-4CAF50?style=for-the-badge\"){kind=icon}
\n Why ·\n Quick start ·\n Tools ·\n Install ·\n Recipes ·\n Pricing ·\n Errors\n
\n\n---\n\n## 🎬 Why\n\nEveryone who works with an agent has run this exchange at least once.\n\n```txt\nYou: Summarize this. https://www.youtube.com/watch?v=kCc8FmEb1nY\nAgent: I'm not able to watch videos. If you paste the transcript here, I'll gladly help!\n```\n\nThe transcript is precisely the thing the agent cannot get on its own. With this server\nconnected, the same message simply resolves.\n\n```txt\nYou: Summarize this. https://www.youtube.com/watch?v=kCc8FmEb1nY\nAgent: → get_transcript(video=\"kCc8FmEb1nY\", video_metadata=true) 1 credit\n\n That's \"Let's build GPT: from scratch, in code, spelled out\" by Andrej\n Karpathy, 1:56:20. He starts from an empty file and a bigram model,\n derives self-attention step by step, and ends with a working GPT that...\n```\n\nReading one video is rarely where the job ends. Here is how the three ways of getting YouTube\ndata into an agent actually compare.\n\n| | This server | Local yt-dlp / scraper MCP | Google YouTube Data API |\n| ------------------------ | ----------- | -------------------------- | ----------------------- |\n| Transcripts | ✅ any public video, 5 formats | ⚠️ blocked on datacenter IPs, breaks when YouTube changes markup | ❌ not served at all |\n| Setup | ✅ a URL and an API key | ❌ local install, binaries to keep alive | ❌ Cloud project, OAuth consent screens |\n| YouTube search | ✅ native, 1 credit per page | ❌ | ⚠️ 100 quota units per search |\n| Channels & playlists | ✅ 100 videos/page, or 500 bare IDs | ❌ one video at a time | ⚠️ quota-metered per item |\n| Bulk transcripts | ✅ 4,000 per background job | ❌ | ❌ |\n| RAG-ready chunking | ✅ 20-5,000 chars, word-level timestamps | ❌ | ❌ |\n| When YouTube changes | ✅ fixed server-side, nothing to update | ❌ you patch and redeploy | ✅ |\n| Failed calls | ✅ credits refund themselves | ❌ your retry logic | ⚠️ quota spent anyway |\n\n---\n\n## ⚡ Quick start\n\n**1. Get an API key.** Sign up at [transcriptout.com](https://transcriptout.com) and create a key\nin the [dashboard](https://transcriptout.com/dashboard). New accounts receive 100 free credits and\nno card is asked. Keys start with `sk_` and are shown once.\n\n**2. Point your client at the server.** It speaks streamable HTTP and authenticates with one\nBearer header.\n\n```json\n{\n \"mcpServers\": {\n \"transcriptout\": {\n \"url\": \"https://api.transcriptout.com/mcp\",\n \"headers\": { \"Authorization\": \"Bearer YOUR_API_KEY\" }\n }\n }\n}\n```\n\nOne-click buttons for Cursor and VS Code sit at the top of this page. Exact snippets for the rest\nlive under [Install in your client](#-install-in-your-client).\n\n**3. Paste a link.**\n\n```txt\nSummarize this talk and pull the three strongest quotes.\nhttps://www.youtube.com/watch?v=dQw4w9WgXcQ\n```\n\nThe agent picks `get_transcript` on its own, reads the timed text and answers from it. Every\nresponse carries an `X-Credits-Remaining` header, so the budget stays in view the whole session.\n\n---\n\n## 🧰 The 14 tools\n\nAll 14 tools are exposed automatically once you connect. **Most calls cost 1 credit.** Credits\nare refunded automatically when a call fails before reaching YouTube (validation errors, rate\nlimits, our own capacity), so you pay for answers, not for failures. A definitive \"this video has no\ncaptions\" is an answer and is billed like one.\n\n### 1. `get_transcript` · 1 credit\n\nFetch the transcript of any YouTube video. `format=text` (default) returns plain readable text,\ncheapest for a model to reason over, and `format=json` returns timed segments.\n\n| Parameter | Type | Default | Description |\n| ---------------- | ------- | ------------ | -------------------------------------------------------------- |\n| `video` | string | **required** | YouTube URL (full or short) or 11-char video ID |\n| `lang` | string | `\"en\"` | Language code of the track (`en`, `de`, ...) |\n| `format` | string | `\"text\"` | `\"text\"` (plain text), `\"json\"` (segments with `start`/`duration` in seconds), `\"srt\"`/`\"vtt\"` (subtitle file body), `\"srv3\"` (raw YouTube XML) |\n| `kind` | string | auto-detect | `\"manual\"` or `\"auto\"`. Left out, a manual track wins and auto is the fallback |\n| `segment` | integer | see below | Max characters per segment. 500-1500 makes RAG-ready chunks |\n| `video_metadata` | boolean | `false` | Add title, channel, duration and views in the SAME call, same 1 credit |\n\nLeft out, `segment` cuts auto-generated tracks into ~180-character segments and returns manual\ntracks exactly as their author broke them. Pass it whenever you need one size regardless of which\ntrack answers.\n\n**Example output (`format=json`):**\n\n```json\n{\n \"video_id\": \"dQw4w9WgXcQ\",\n \"language\": \"en\",\n \"kind\": \"manual\",\n \"transcript\": [\n { \"text\": \"Never gonna give you up\", \"start\": 18.0, \"duration\": 4.12 },\n { \"text\": \"Never gonna let you down\", \"start\": 22.12, \"duration\": 3.85 }\n ]\n}\n```\n\n> `srt` and `vtt` come back as complete subtitle file bodies, ready to be written to disk by the\n> agent. `srv3` is the raw source XML and does not combine with `segment`.\n\n### 2. `get_video_info` · 1 credit\n\nMetadata for one video (title, channel, duration, views, thumbnails) plus the list of available\ntranscript languages, WITHOUT downloading the subtitles.\n\n| Parameter | Type | Default | Description |\n| --------- | ------ | ------------ | -------------------------- |\n| `id` | string | **required** | YouTube video ID or URL |\n\n> **Credit hygiene:** if you are going to fetch the transcript anyway, call `get_transcript` with\n> `video_metadata=true` instead. It returns both for one credit where these are two calls and two.\n\n### 3. `search_youtube` · 1 credit/page\n\nSearch YouTube for videos or channels. Paginate with `next_page_token`. `has_more` tells you\nwhether another page exists.\n\n| Parameter | Type | Default | Description |\n| ----------------- | ------- | --------- | ------------------------------------ |\n| `q` | string | required* | Search query (*unless paginating) |\n| `type` | string | `\"video\"` | `\"video\"` or `\"channel\"` |\n| `limit` | integer | `20` | Results per page, 1-50 |\n| `next_page_token` | string | | Token from a previous result |\n\n### 4. `list_channel_videos` · 1 credit/page\n\nList videos from a channel's Videos tab, newest first. Accepts an `@handle`, a channel name, a\n`UC...` channel ID or a channel URL.\n\n| Parameter | Type | Default | Description |\n| ----------------- | ------- | --------- | -------------------------------------------------- |\n| `name` | string | required* | `@handle`, channel name, `UC...` ID or URL |\n| `limit` | integer | `100` | Page size, up to 500 with `ids_only` |\n| `ids_only` | boolean | `false` | Return just `video_ids[]`, up to 500 per page |\n| `next_page_token` | string | | Token from a previous result |\n\n> `ids_only=true` is the cheap way to feed `submit_transcripts_job`.\n\n### 5. `search_channel_videos` · 1 credit/page\n\nSearch inside one channel using YouTube's native relevance search. A result whose title lacks the\nquery word is normal. Results are ranked by relevance, not by substring.\n\n| Parameter | Type | Default | Description |\n| ----------------- | ------- | --------- | ------------------------------------ |\n| `name` | string | required | `@handle`, channel name, `UC...` ID or URL |\n| `q` | string | required | Query to search within the channel |\n| `limit` | integer | `30` | Results per page, 1-100 |\n| `next_page_token` | string | | Pagination token |\n\n### 6. `latest_channel_videos` · 1 credit\n\nThe ~15 most recent videos of a channel from its RSS feed. The fastest and cheapest way to check\nwhat a channel published recently.\n\n| Parameter | Type | Default | Description |\n| --------- | ------ | ------------ | ------------------------------------------ |\n| `name` | string | **required** | `@handle`, channel name, `UC...` ID or URL |\n\n### 7. `list_playlist_videos` · 1 credit/page\n\nEvery video of a playlist in playlist order. Accepts a `PL...` playlist ID or a URL with `list=`.\n\n| Parameter | Type | Default | Description |\n| ----------------- | ------- | --------- | --------------------------------------------- |\n| `id` | string | required* | Playlist ID or URL |\n| `limit` | integer | `100` | Page size, up to 500 with `ids_only` |\n| `ids_only` | boolean | `false` | Return just `video_ids[]`, up to 500 per page |\n| `next_page_token` | string | | Pagination token |\n\n### 8. `search_playlist_videos` · 1 credit\n\nFind videos inside a playlist by a substring of the title (case-insensitive). YouTube has no native\nplaylist search, so this scans up to 500 playlist items. `truncated=true` means there may be more\nmatches beyond the scanned window.\n\n| Parameter | Type | Default | Description |\n| --------- | ------- | -------- | ---------------------------------- |\n| `id` | string | required | Playlist ID or URL |\n| `q` | string | required | Substring to match in video titles |\n| `limit` | integer | `30` | Max matches, 1-100 |\n\n### 9. `submit_transcripts_job` · 1 credit per video\n\nQueue transcripts for MANY videos at once (up to 4,000) and get a `job_id` back immediately. The\nwork continues in the background at your rate limit's pace. Use this instead of calling\n`get_transcript` in a loop for more than a handful of videos.\n\n| Parameter | Type | Default | Description |\n| ----------------- | -------- | ------------ | ---------------------------------------------------- |\n| `videos` | string[] | **required** | Video IDs or URLs, up to 4,000. Duplicates collapse BEFORE billing |\n| `lang` | string | `\"en\"` | One language for the whole job |\n| `format` | string | `\"text\"` | `\"text\"`, `\"json\"`, `\"srt\"`, `\"vtt\"` or `\"srv3\"`, one for the whole job |\n| `kind` | string | auto-detect | `\"manual\"` or `\"auto\"` |\n| `segment` | integer | | One segment size for the whole job |\n| `video_metadata` | boolean | `false` | Metadata per video, no extra cost |\n| `idempotency_key` | string | | Resubmitting the same list with the same key returns the SAME job, no double charge |\n\nRequires a user key (`sk_...`). Credits are charged on submit and refunded per video when a video\ncould not be delivered through our fault.\n\n### 10. `get_transcripts_job` · free\n\nProgress of a batch job: status (`queued`/`running`/`done`/`cancelled`), how many videos are ready,\nfailed and pending. Polling a job you already paid for costs nothing.\n\n### 11. `get_transcripts_results` · free\n\nFinished transcripts from a batch job, in the order submitted, paged with `next_page_token`\n(`limit` 1-500, default 100). Results appear as they are fetched, so you can read before the job is\ndone. Each entry is exactly what `get_transcript` returns for that video, plus its status.\n\n### 12. `get_transcripts_result` · free\n\nOne video's result out of a batch job, by its video id, without paging through the whole result\nset. A 404 means the job does not exist or this video has not finished yet, so check\n`get_transcripts_job` before concluding anything.\n\n| Parameter | Type | Default | Description |\n| ---------- | ------ | ------------ | ---------------------------------------------- |\n| `job_id` | string | **required** | Job id from `submit_transcripts_job` |\n| `video_id` | string | **required** | One of the video ids the job was submitted with |\n\n### 13. `cancel_transcripts_job` · free\n\nCancel a batch job. Credits are refunded ONLY for videos not started yet. Anything already fetched\nstays in the results and stays paid for.\n\n### 14. `get_credits` · free\n\nThe remaining credit balance of the key, with no parameters. The balance also rides in the\n`X-Credits-Remaining` header of every response, but headers are invisible to the model, so the\nnumber a user actually asks about needs a tool. Handy right before a large batch too, since the\nbatch charges 1 credit per video on submit.\n\n---\n\n## 🔌 Install in your client\n\nThe server is remote, so every install below is a config entry and nothing more. All of them want the same two values, the URL and the Bearer header from [Quick start](#-quick-start).\n\n> **Worth doing once, a standing rule for your client**\n>\n> With this in your client's rules/instructions, pasting a YouTube link is enough and the word\n> \"transcript\" never has to be typed:\n>\n> ```txt\n> Whenever a YouTube link or video ID appears in my message, call the\n> transcriptout get_transcript tool first and answer from the transcript,\n> whether I asked for a summary, a quote, a translation or a question.\n> ```\n\n![\"Install](\"https://img.shields.io/badge/VS_Code-Install_TranscriptOut_MCP-0098FF?style=for-the-badge&logo=visualstudiocode&logoColor=white\"){kind=icon}
](https://insiders.vscode.dev/redirect?url=vscode%3Amcp%2Finstall%3F%7B%22name%22%3A%22transcriptout%22%2C%22url%22%3A%22https%3A%2F%2Fapi.transcriptout.com%2Fmcp%22%7D)\n\nOr add this to VS Code user settings (`settings.json`):\n\n```json\n\"mcp.servers\": {\n \"transcriptout\": {\n \"type\": \"http\",\n \"url\": \"https://api.transcriptout.com/mcp\",\n \"headers\": {\n \"Authorization\": \"Bearer YOUR_API_KEY\"\n }\n }\n}\n```\n\n\n © 2026 TranscriptOut · Released under the MIT License\n
\n", "bytes": 32072, "sha": "39e7f344e6f039147fac10bab8f4812319df57a9f01eda67d97ee6d130d1ffcf", "repo_slug": "artemchuikin/youtube-mcp", "fonte": "repo", "truncated": false, "api": "https://agentalog.com/api/listings/mcp_com_transcriptout_youtube_transcript_and_2bde71f6/readme" }