About TubeToTranscript
TubeToTranscript is an MCP server that retrieves captions from YouTube videos for AI assistants and other MCP clients. It provides timed transcripts, available caption languages, and exports in text and subtitle formats.
Key features
- 01Fetches timed video captions
- 02Lists available caption languages
- 03Exports TXT, timestamped text, SRT, and VTT
- 04Accepts video links or IDs
Use cases
- 01Summarizing a captioned YouTube video
- 02Finding a quote and its timestamp
- 03Creating study notes from a lecture
- 04Exporting captions as subtitles
MCP support
The MCP server exposes tools to retrieve YouTube transcripts, list caption languages, and export transcripts. It can run locally over stdio or connect to the remote endpoint over Streamable HTTP.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://www.tubetotranscript.com/mcpclaude mcp add --transport http tubetotranscript-tubetotranscript https://www.tubetotranscript.com/mcp{"mcpServers":{"tubetotranscript-tubetotranscript":{"type":"http","url":"https://www.tubetotranscript.com/mcp"}}}Open in Cursor{"servers":{"tubetotranscript-tubetotranscript":{"type":"http","url":"https://www.tubetotranscript.com/mcp"}}}Open in VS Code{"mcpServers":{"tubetotranscript-tubetotranscript":{"type":"http","url":"https://www.tubetotranscript.com/mcp"}}}{"mcpServers":{"tubetotranscript-tubetotranscript":{"serverUrl":"https://www.tubetotranscript.com/mcp"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
https://www.tubetotranscript.com/mcpAuth requiredTools
Details
From the project READMETubeToTranscript MCP Server: YouTube Transcripts for Claude, Cursor & AI Agents
By TubeToTranscript, the free YouTube transcript generator. Website · API docs · Remote MCP endpoint · npm · Source
tubetotranscript-mcp is a Model Context Protocol (MCP) server
that lets AI assistants read YouTube videos. Ask Claude, Cursor, or any MCP client to summarize a
video, find a quote, or turn a lecture into notes, and it fetches the video's captions itself, with
timestamps, title, channel, and language.
- Free, no API key, no account. Runs on your computer and reads the captions YouTube already has.
- Works with any video that has captions: watch links, Shorts, youtu.be links, embeds, or a bare video ID.
- Every caption language the video offers, including auto-generated tracks.
- Exports plain text, text with timestamps, SRT, and WebVTT subtitles.
Quick start
Point your MCP client at the package with npx. Nothing to install globally.
Claude Desktop
Add to claude_desktop_config.json
(macOS: ~/Library/Application Support/Claude/, Windows: %APPDATA%\Claude\):
{
"mcpServers": {
"tubetotranscript": {
"command": "npx",
"args": ["-y", "tubetotranscript-mcp"]
}
}
}
Cursor
Add to .cursor/mcp.json:
{
"mcpServers": {
"tubetotranscript": {
"command": "npx",
"args": ["-y", "tubetotranscript-mcp"]
}
}
}
Restart the client afterwards so it picks up the new server.
Claude Code
claude mcp add tubetotranscript -- npx -y tubetotranscript-mcp
What you can ask
Once connected, just paste a YouTube link into the chat:
- "Summarize this video in five bullet points: https://youtu.be/…"
- "When does the speaker talk about pricing? Give me the timestamps."
- "Turn this lecture into study notes and ten flashcards."
- "Which caption languages does this video have?"
- "Export the Spanish captions as an SRT file."
Tools
get_youtube_transcript
Returns the transcript with timed segments, title, channel, duration, and the available languages.
| Parameter | Required | Description |
|---|---|---|
url_or_video_id | yes | Video URL (watch, shorts, youtu.be, embed) or the 11-character ID |
language | no | Language code such as en, es, fr. Defaults to English or the primary track |
get_available_languages
Lists every caption and subtitle track available for a video.
| Parameter | Required | Description |
|---|---|---|
url_or_video_id | yes | Video URL or 11-character ID |
export_youtube_transcript
Converts a transcript into a text or subtitle format.
| Parameter | Required | Description |
|---|---|---|
url_or_video_id | yes | Video URL or 11-character ID |
format | yes | txt, txt_with_timestamps, srt, or vtt |
language | no | Language code |
Local or remote?
| Local (this package) | Remote (tubetotranscript.com/mcp) | |
|---|---|---|
| Install | npx -y tubetotranscript-mcp | None, connect by URL |
| Transport | stdio | Streamable HTTP |
| API key | Not needed | Required (get a free key) |
| Where requests run | Your computer | TubeToTranscript's servers |
Clients that take a remote URL, such as Cursor:
{
"mcpServers": {
"tubetotranscript": {
"url": "https://www.tubetotranscript.com/mcp",
"headers": { "Authorization": "Bearer YOUR_API_KEY" }
}
}
}
Without an AI client
The same transcripts are available on the web at TubeToTranscript, free and without signing up:
- YouTube Transcript Generator: paste a link, read, search, copy, or download
- YouTube Transcript for ChatGPT & Claude: transcript plus prompt, split for long videos
- YouTube Quote Finder: find when something was said
- YouTube to JSON and YouTube to Excel & CSV exports
- REST API: the same data over HTTP for scripts and apps
Requirements & limitations
- Node.js 18 or later (uses the built-in
fetch). - Reads only captions that YouTube already exposes. Videos with captions turned off return an error; this server does no speech-to-text of its own.
- Caption text can contain recognition errors. Check names, figures, and quotes against the video.
License
MIT. Made by TubeToTranscript.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.