Youtube Research Mcp

@lee-s-dev

About Youtube Research Mcp

YouTube MCP server — transcripts, comments & channel analysis as AI research source

Config

Add this server to your MCP-compatible client using the configuration below.

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

Tools

No tools detected

Fetch the live tool list by running this server in a temporary sandbox using the button above.

Overview

What is Youtube Research Mcp?

A server that integrates YouTube subtitles, comments, and channel data as research sources for AI assistants via the Model Context Protocol (MCP). It works with any MCP‑capable AI client (Claude Desktop, Cursor, Windsurf, Cline, etc.) and requires no API key for subtitle and channel analysis; an optional YouTube Data API key unlocks video search and comment analysis.

How to use Youtube Research Mcp?

Install it with uvx youtube-research-mcp (recommended, requires uv) or pip install youtube-research-mcp. Then configure your MCP client (e.g., Claude Desktop’s claude_desktop_config.json) with the command uvx and arguments ["youtube-research-mcp"]. To use all features, set the environment variable YOUTUBE_API_KEY. Restart the client after configuration.

Key features of Youtube Research Mcp

Fetch video subtitles without any API key.
Analyze multiple videos in parallel (subtitles + optionally comments).
Analyze a YouTube channel’s recent videos.
Search videos by keyword (requires API key).
Collect video comments and discussions (requires API key).
Cache results in local SQLite to avoid redundant API calls.
Prompt injection defense via a safety_notice field.
No external data transmission or LLM calls.

Use cases of Youtube Research Mcp

Summarize the key points of a single YouTube video.
Compare arguments, commonalities, and differences across multiple videos.
Analyze recent content trends of a specific YouTube channel.
Search for videos on a topic and collect their subtitles and comments for research.
Compare creator claims with real audience reactions (comments).

FAQ from Youtube Research Mcp

Do I need a YouTube API key?

No, for fetching subtitles and channel metadata. A free API key is required only for video search and comment collection (daily quota of 10,000 units).

Is it safe to install and use?

Yes. The server only fetches data from YouTube and caches it locally. It never sends collected data to external servers, never calls any LLM or AI API, never accesses files outside its cache directory, and never executes shell commands. Source code is public on GitHub.

What data is cached and for how long?

Subtitles are cached for 30 days, comments for 6 hours, search results for 2 hours, and video metadata permanently. Caching reduces duplicate API calls.

Where is the cache stored?

Default locations: macOS ~/Library/Application Support/youtube-research-mcp/cache.db, Windows %APPDATA%\youtube-research-mcp\cache.db, Linux ~/.local/share/youtube-research-mcp/cache.db. Override with the YOUTUBE_RESEARCH_CACHE_DB environment variable.

How do I install Youtube Research Mcp?

Use uvx youtube-research-mcp (recommended, requires uv) or pip install youtube-research-mcp. For pip installations, set the MCP client’s command to "youtube-research-mcp" with empty args.

Frequently asked questions

Do I need a YouTube API key?

No, for fetching subtitles and channel metadata. A free API key is required only for video search and comment collection (daily quota of 10,000 units).

Is it safe to install and use?

What data is cached and for how long?

Subtitles are cached for 30 days, comments for 6 hours, search results for 2 hours, and video metadata permanently. Caching reduces duplicate API calls.

Where is the cache stored?

Default locations: macOS `~/Library/Application Support/youtube-research-mcp/cache.db`, Windows `%APPDATA%\youtube-research-mcp\cache.db`, Linux `~/.local/share/youtube-research-mcp/cache.db`. Override with the `YOUTUBE_RESEARCH_CACHE_DB` environment variable.

How do I install Youtube Research Mcp?

Use `uvx youtube-research-mcp` (recommended, requires `uv`) or `pip install youtube-research-mcp`. For pip installations, set the MCP client’s command to `"youtube-research-mcp"` with empty `args`.

Comments

Basic information

More Data & Analytics MCP servers

TofuBofu AI Visibility

Arnav Neil

TofuBofu runs a free AI visibility scan for any B2B company, right inside your AI assistant. Ask it to scan a domain and it runs the questions real buyers ask across ChatGPT, Claude, Gemini, Perplexity, Google AI Overvie

cloro

Extract structured data from ChatGPT, Perplexity, Grok, Gemini, Google Search, Google News, Copilot, and AI Overview through one API, across all countries and states.

LeadMarina

Verified local-business leads: search any niche in any city, query your lead library, and export to Close, GoHighLevel, Google Sheets, or emailed files — every email SMTP-checked, every phone verified with line type and

Octolens

Give your AI agents every brand mention across the internet — 15+ platforms, AI-filtered, one structured feed.

Bounce Watch - Company Signal Intelligence

bouncewatch

Bounce Watch tells you what just changed at a company, and when. Who raised money. Who hired a senior person. Who opened an office, won a customer, or announced a partnership. Every event carries its own date, so you ca