YouTube Transcript MCP Server (Jkawamoto)

YouTube transcripts and video metadata retrieval for AI agents.

Data last scanned today · Reviewed yesterday

454 tools★ 458PythonMIT

Overview

This server enables AI models to directly access YouTube content, specifically transcriptions and video metadata. It provides tools for fetching full transcripts, obtaining timed transcripts with associated timestamps, and querying specific language availability for video content. Integration is supported through uv for standard environments or via direct .mcpb bundle installation for Claude Desktop users. It requires a valid YouTube video URL for all operations, making it a utility for synthesizing information from long-form video content into an AI-accessible format.

Our verdict

This server carries a high risk level with a security score of 30/100. The primary concern is the total lack of an authentication method, meaning any local configuration access allows for server use. Given that it also lacks a read-only mode and provides full write access by default, users should exercise caution. It is suitable only for trusted, local environments where users can manage the associated risks of unauthenticated execution.

Zero direct dependencies minimize supply-chain surface area.
Active maintenance with commits within the last ten days.
Supports fetching metadata and multi-language transcripts.

No authentication method, allowing unrestricted access if configured.
Lacks a read-only mode, defaulting to full write access.
Community-maintained project without official vendor backing.

Tools

Tool	Description	Risk
get_transcript	Fetches the transcript of a specified YouTube video.	low
get_timed_transcript	Fetches the transcript of a specified YouTube video with timestamps.	low
get_video_info	Fetches the metadata of a specified YouTube video.	low
get_available_languages	Retrieves the available languages for the video.	low

Compatibility

Client	Local	Docker	Remote	Read-only
Claude Desktop
Cursor
VS Code
Windsurf
ChatGPT

Frequently asked questions

›What prerequisites are required to use this server?

You must have the uv package manager installed on your system to use the server via the provided uvx commands.

›How can I manage long transcripts that exceed token limits?

The server automatically splits transcripts exceeding 50,000 characters and provides a next_cursor in the response for pagination. You can customize this threshold by using the --response-limit command line argument.

›Does the server support multiple languages?

Yes, you can specify a language code using the lang parameter, which defaults to 'en'. You can use the get_available_languages tool to see which languages are supported for a specific video.

›Can I use this server behind a proxy?

Yes, you can configure proxy settings using environment variables like HTTP_PROXY or HTTPS_PROXY, or by using command line arguments for specific proxies like Webshare.

›Are there different tools for transcripts with and without timestamps?

Yes, the server provides separate tools: get_transcript for raw text and get_timed_transcript for transcripts that include timestamp information.