varuniJ/Podcast-Summarizer-MCP

Podcast Summarizer MCP

A Model Context Protocol (MCP) server that downloads YouTube videos, transcribes them using Whisper, and summarizes them using Groq's LLM.

Setup

1. Install dependencies:

pip install -r requirements.txt

2. Set up your environment variables:

# Create a .env file with your Groq API key
echo "GROQ_API_KEY=your_groq_api_key_here" > .env

Usage

Option 1: Run Server and Client Separately

1. Start the MCP server:

python podcast_mcp_server.py

2. In another terminal, run the client:

python mcp_client.py

Option 2: Use with Claude Desktop

1. Start the MCP server:

python podcast_mcp_server.py

2. In Claude Desktop, add the MCP server:
   • Go to Settings → Model Context Protocol
   • Add a new server
   • Command: python podcast_mcp_server.py
   • Working directory: Your project folder

How it Works

1. Download: Uses yt-dlp to download audio from YouTube URLs
2. Transcribe: Uses faster-whisper to convert audio to text
3. Summarize: Uses Groq's llama3-70b-8192 model to create a 5-point summary

Available Tools

• summarize_podcast(url: str): Downloads, transcribes, and summarizes a YouTube video

Example

🎧 Enter YouTube URL to summarize (or 'quit' to exit): https://www.youtube.com/watch?v=example

⏳ Summarizing: https://www.youtube.com/watch?v=example
This may take a few minutes...

==================================================
📝 SUMMARY RESULT:
==================================================
✅ **Summary of the Podcast**  
Language: `en`

**Summary:**  
1. Point one about the content
2. Point two about the content
3. Point three about the content
4. Point four about the content
5. Point five about the content
==================================================

Requirements

• Python 3.8+
• FFmpeg (for audio processing)
• Groq API key
• Internet connection for downloading videos