YouTube Transcript Extractor with MCP

A web application that extracts transcripts from YouTube videos using the Model Context Protocol (MCP) for communication between components.

Architecture

graph TB
    subgraph Frontend
        UI[Gradio Web Interface]
        Client[MCP Client]
    end

    subgraph Backend
        Server[MCP Server]
        YT[YouTube Transcript API]
    end

    UI --> |User Input| Client
    Client --> |MCP Protocol| Server
    Server --> |API Calls| YT
    YT --> |Transcript Data| Server
    Server --> |MCP Response| Client
    Client --> |Display| UI

    style UI fill:#f9f,stroke:#333,stroke-width:2px
    style Server fill:#bbf,stroke:#333,stroke-width:2px
    style YT fill:#bfb,stroke:#333,stroke-width:2px

Component Flow

sequenceDiagram
    participant User
    participant UI as Gradio Interface
    participant Client as MCP Client
    participant Server as MCP Server
    participant YT as YouTube API

    User->>UI: Enter YouTube URL
    UI->>Client: Request Transcript
    Client->>Server: Initialize Connection
    Server->>Client: Connection Established
    Client->>Server: Call get_transcript Tool
    Server->>YT: Request Transcript
    YT->>Server: Return Transcript Data
    Server->>Client: Return Formatted Text
    Client->>UI: Display Transcript
    UI->>User: Show Result

Features

• Extract transcripts from any YouTube video
• Support for multiple languages (falls back to English if requested language isn't available)
• Optional timestamp inclusion
• Clean and intuitive web interface
• Robust error handling and fallbacks
• Efficient client-server communication using MCP

Setup

1. Clone the repository:

git clone <repository-url>
cd youtube-mcp

2. Create and activate a Python virtual environment:

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies:

pip install -r requirements.txt

Usage

1. Start the application:

python gradio_youtube_transcript.py

2. Open your web browser and navigate to the provided URL (typically http://localhost:7860)

3. Enter a YouTube URL and select your preferences:

   • Choose the desired language
   • Toggle timestamps on/off
   • Click "Get Transcript" to extract

Technical Details

Components

1. Gradio Interface (gradio_youtube_transcript.py)

   • Provides the web user interface
   • Handles user input and display
   • Manages client-server communication

2. MCP Server (yt_dlp_transcript_server.py)

   • Implements the Model Context Protocol
   • Provides transcript extraction tools
   • Handles YouTube API interaction
   • Manages error handling and fallbacks

Tools

The MCP server provides two main tools:

1. get_transcript

   • Gets plain text transcript
   • Parameters:
     • video_url: YouTube video URL
     • language: Preferred language code (defaults to 'en')

2. get_transcript_with_timestamps

   • Gets transcript with timestamps
   • Parameters:
     • video_url: YouTube video URL
     • language: Preferred language code (defaults to 'en')

Communication Protocol

The application uses the Model Context Protocol (MCP) for client-server communication:

flowchart LR
    subgraph Protocol
        direction TB
        Init[Initialize Connection]
        List[List Available Tools]
        Call[Call Tool]
        Resp[Process Response]
    end

    Init --> List
    List --> Call
    Call --> Resp
    Resp --> |New Request| Call

    style Protocol fill:#f5f5f5,stroke:#333,stroke-width:2px

Error Handling

The application implements robust error handling:

1. Connection Errors

   • Automatic reconnection attempts
   • Clear error messages to users

2. Transcript Errors

   • Language fallback system
   • Informative error messages

3. Resource Management

   • Proper cleanup of connections
   • Memory leak prevention

Requirements

• Python 3.8+
• Dependencies listed in requirements.txt:
  • mcp>=0.1.0
  • youtube-transcript-api>=0.6.1
  • gradio>=4.19.2

License

[Add your license information here]