What Is YouTube MCP Server?

YouTube MCP Server is a Model Context Protocol server that lets Claude search YouTube, pull video transcripts, and analyze channel content without leaving your conversation. It is maintained by Zubeid Hendricks and published under the mcp-youtube package on npm.

Quick Facts

Field	Value
Package	`mcp-youtube`
Install	`npx mcp-youtube`
Compatible with	Claude Desktop, Claude Code
Status	active
GitHub	https://github.com/ZubeidHendricks/youtube-mcp-server

What You Can Do With It

Video search: Query YouTube and retrieve video results directly from within a Claude conversation.
Transcript extraction: Pull the full spoken-word transcript from a given video URL, useful for summarization or citation work.
Channel analysis: Inspect a YouTube channel’s content to review uploads, topics, or posting patterns.
Content research: Combine search and transcript tools to gather source material across multiple videos in a single workflow.

How to Set It Up

Prerequisites: Node.js 18 or later must be installed.

Step 1 — Test the server runs locally

npx mcp-youtube
**Step 2 — Add it to Claude Desktop**

Open or create `~/Library/Application Support/Claude/claude_desktop_config.json` (macOS) or the equivalent path on Windows, and add the following entry under `mcpServers`:

```json
{
  "mcpServers": {
    "youtube": {
      "command": "npx",
      "args": ["mcp-youtube"]
    }
  }
}

Restart Claude Desktop after saving.

Step 3 — Add it to Claude Code

In your project root, open or create .claude/mcp.json and add:

{
  "mcpServers": {
    "youtube": {
      "command": "npx",
      "args": ["mcp-youtube"]
    }
  }
}

Run claude mcp list to confirm the server is registered.

Works Best With

YouTube MCP Server is documented as compatible with both Claude Desktop and Claude Code. Claude Desktop is the more common pairing for on-demand research workflows — searching a topic, extracting a transcript, and summarizing in one session. Claude Code is the right choice when transcript data or channel metadata needs to feed into a script, file, or automated pipeline within a development project.

Known Limitations

YouTube Data API quota: If the server uses the YouTube Data API v3 under the hood, Google enforces a default daily quota of 10,000 units. Bulk transcript or search operations can exhaust this quota quickly. Check the repository’s environment variable documentation for API key configuration requirements.
Transcript availability: Not all videos have transcripts. Auto-generated captions may be absent for non-English content or very recent uploads. The server can only return what YouTube exposes.
No video playback or download: This server retrieves metadata and text data only — it does not stream, download, or process audio or video files.
Rate limiting: npx fetches the package on each invocation if not cached; consider installing globally (npm install -g mcp-youtube) for faster startup in repeated use.
Auth requirements: Review the GitHub repository for any required API keys before assuming unauthenticated access works in all environments.