Skip to main content
Glama

YouTube Vision MCP Server (youtube-vision)

NPM version License: MIT

MCP (Model Context Protocol) server that utilizes the Google Gemini Vision API to interact with YouTube videos. It allows users to get descriptions, summaries, answers to questions, and extract key moments from YouTube videos.

Features

  • Analyzes YouTube videos using the Gemini Vision API.

  • Provides multiple tools for different interactions:

    • General description or Q&A (ask_about_youtube_video)

    • Summarization (summarize_youtube_video)

    • Key moment extraction (extract_key_moments)

  • Lists available Gemini models supporting generateContent.

  • Configurable Gemini model via environment variable.

  • Communicates via stdio (standard input/output).

Related MCP server: YouTube Insights MCP Server

Prerequisites

Before using this server, ensure you have the following:

  • Node.js: Version 18 or higher recommended. You can download it from nodejs.org.

  • Google Gemini API Key: Obtain your API key from Google AI Studio or Google Cloud Console.

Installation & Usage

There are two main ways to use this server:

Installing via Smithery

To install youtube-vision-mcp for Claude Desktop automatically via Smithery:

npx -y @smithery/cli install @minbang930/youtube-vision-mcp --client claude

The easiest way to run this server is using npx, which downloads and runs the package without needing a permanent installation.

You can configure it within your MCP client's settings file (Claude, VSCode .. ):

{
  "mcpServers": {
    "youtube-vision": {
      "command": "npx",
      "args": [
        "-y",
        "youtube-vision"
      ],
      "env": {
        "GEMINI_API_KEY": "YOUR_GEMINI_API_KEY",
        "GEMINI_MODEL_NAME": "gemini-2.0-flash"
      }
    }
  }
}

Replace "YOUR_GEMINI_API_KEY" with your actual Google Gemini API key.

Option 2: Manual Installation (from Source)

If you want to modify the code or run it directly from the source:

  1. Clone the repository:

    git clone https://github.com/minbang930/Youtube-Vision-MCP.git
    cd youtube-vision
  2. Install dependencies:

    npm install
  3. Build the project:

    npm run build
  4. Configure and run: You can then run the compiled code using node dist/index.js directly (ensure GEMINI_API_KEY is set as an environment variable) or configure your MCP client to run it using the node command and the absolute path to dist/index.js, passing the API key via the env setting as shown in the npx example.

Configuration

The server uses the following environment variables:

  • GEMINI_API_KEY (Required): Your Google Gemini API key.

  • GEMINI_MODEL_NAME (Optional): The specific Gemini model to use (e.g., gemini-1.5-flash). Defaults to gemini-2.0-flash. Important: For production or commercial use, ensure you select a model version that is not marked as "Experimental" or "Preview".

Environment variables should be set in the env section of your MCP client's settings file (e.g., mcp_settings.json).

Available Tools

1. ask_about_youtube_video

Answers a question about the video or provides a general description if no question is asked.

  • Input:

    • youtube_url (string, required): The URL of the YouTube video.

    • question (string, optional): The specific question to ask about the video. If omitted, a general description is generated.

  • Output: Text containing the answer or description.

2. summarize_youtube_video

Generates a summary of a given YouTube video.

  • Input:

    • youtube_url (string, required): The URL of the YouTube video.

    • summary_length (string, optional): Desired summary length ('short', 'medium', 'long'). Defaults to 'medium'.

  • Output: Text containing the video summary.

3. extract_key_moments

Extracts key moments (timestamps and descriptions) from a given YouTube video.

  • Input:

    • youtube_url (string, required): The URL of the YouTube video.

    • number_of_moments (integer, optional): Number of key moments to extract. Defaults to 3.

  • Output: Text describing the key moments with timestamps.

4. list_supported_models

Lists available Gemini models that support the generateContent method (fetched via REST API).

  • Input: None

  • Output: Text listing the supported model names.

Important Notes

  • Model Selection for Production: When using this server for production or commercial purposes, please ensure the selected GEMINI_MODEL_NAME is a stable version suitable for production use. According to the Gemini API Terms of Service, models marked as "Experimental" or "Preview" are not permitted for production deployment.

  • API Terms of Service: Usage of this server relies on the Google Gemini API. Users are responsible for reviewing and complying with the Google APIs Terms of Service and the Gemini API Additional Terms of Service. Note that data usage policies may differ between free and paid tiers of the Gemini API. Do not submit sensitive or confidential information when using free tiers.

  • Content Responsibility: The accuracy and appropriateness of content generated via the Gemini API are not guaranteed. Use discretion before relying on or publishing generated content.

License

This project is licensed under the MIT License. See the LICENSE file for details.

Available Tools

4 tools
ask_about_youtube_videoB

Answers a question about the video or provides a general description if no question is asked.

Inputs: youtube_url (required), question.

extract_key_momentsC

Extracts key moments (timestamps and descriptions) from a given YouTube video.

Inputs: youtube_url (required), number_of_moments.

list_supported_modelsB

Lists available Gemini models that support the 'generateContent' method.

Inputs: no named parameters.

summarize_youtube_videoC

Generates a summary of a given YouTube video URL using Gemini Vision API.

Inputs: youtube_url (required), summary_length.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    B3.1/5.0

    Scored across 4 tools

    Disambiguation4/5

    The tools have mostly distinct purposes, but 'ask_about_youtube_video' and 'summarize_youtube_video' could be confused as both provide descriptive outputs about video content. However, 'ask_about_youtube_video' is question-driven while 'summarize_youtube_video' is general, and the other tools ('extract_key_moments', 'list_supported_models') are clearly differentiated.

    Naming Consistency3/5

    Three tools follow a consistent verb_noun pattern ('ask_about_youtube_video', 'extract_key_moments', 'summarize_youtube_video'), but 'list_supported_models' deviates by using 'list' instead of a more descriptive verb like 'get' or 'retrieve', and it lacks the 'youtube_video' domain specificity. This mixed convention reduces predictability.

    Tool Count4/5

    With 4 tools, the count is reasonable for a focused YouTube video analysis server. It covers core functionalities like description, summarization, moment extraction, and model listing, though it could be slightly expanded for more comprehensive coverage (e.g., adding video metadata retrieval).

    Completeness3/5

    The server covers key video analysis tasks (description, summarization, moment extraction) and model support listing, but there are notable gaps. For example, it lacks tools for video metadata (e.g., title, duration, uploader), search capabilities, or interaction with YouTube's API beyond vision-based analysis, which limits agent workflows in broader YouTube contexts.

    Maintenance

    ActivityInactive
    ResponsivenessLimited data

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      A
      quality
      C
      maintenance
      MCP server that fetches YouTube video transcripts and optionally summarizes them. Supports multiple transcript formats (text, JSON, SRT, WebVTT), multi-language retrieval, and flexible YouTube URL parsing.
      6
      49 PyPI
      6
      MIT
    • A
      license
      Not graded
      quality
      D
      maintenance
      A local MCP server for extracting YouTube video transcripts, metadata, and performing visual analysis using Gemini Vision or local Whisper models. It enables users to process video content through various tools for subtitle retrieval and frame analysis.
      23 npm
      MIT