MCP Server

🚀 MCP Server

The MCP Server is built on top of all the APIs of YouTube-Summarizer. All APIs are exposed as tools in the MCP protocol and can be integrated and used by any AI application.

⚠️ Important Note

Currently, MCP only supports local connections, so the use of remote tools is not supported.

🚀 Quick Start

📦 Installation

Docker Installation and Configuration

Build the Docker image:

docker build -t youtube-summarizer-mcp .

Run the MCP server:

docker run -i --rm youtube-summarizer-mcp

Using the Inspector

You can use the MCP Inspector to explore available tools and conduct tests:

./inspector.sh

Integration with Claude Desktop

Add the following content to your claude_desktop_config.json:

{
  "mcpServers": {
    "youtube-summarizer": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "youtube-summarizer-mcp"
      ]
    }
  }
}

Now you can use the MCP tools from server.py in Claude Desktop.

MCP Client Example (without Claude Desktop)

Run the following scripts to set up:

Setup

./setup.sh

Run

./run.sh

It will run both the MCP server and the client and connect them. The terminal will prompt the user to enter a natural language query, which will then be converted into MCP tool calls to answer the user's query.