YouTube Data Access and Analysis Service (MCP Implementation)

This Model Context Protocol (MCP) service furnishes functionalities for probing the YouTube platform, acquiring synchronized text captions, and executing deep semantic pattern matching across the retrieved video transcripts.

Support Channel

Should this utility prove beneficial for your endeavors, kindly consider offering monetary encouragement to sustain ongoing development. Your patronage fuels the creation of sophisticated artificial intelligence instruments.

Your generous backing directly contributes to our momentum in engineering intuitive and potent computational tools.

Structural Outline

• YouTube Data Access and Analysis Service
• Core Capabilities
• Prerequisites for Operation
• Setup Procedure
• Activating the Service Interface
• Method A: Direct Execution
• Method B: Configuration for Claude.app Integration
• Exposed Toolset
• Utilization with MCP Consumers
• Illustrative Example
• Troubleshooting Guide
• Contribution Guidelines
• Licensing Information

Core Capabilities

• Capability to search the YouTube catalog without reliance on the official YouTube Data API.
• Facility for obtaining full video subtitle tracks (transcripts).
• Mechanism for persisting video metadata alongside transcribed text into a vector data repository.
• Functionality for executing conceptual retrieval queries against the indexed transcript content.

Prerequisites for Operation

• Execution environment requires Python version 3.8 or newer.
• A valid Google Cloud Platform API key designated for generating vector embeddings.
• Installation of the uv package manager.

Setup Procedure

1. Duplicate the project repository contents.

2. Establish and invoke a dedicated virtual environment using uv: bash uv venv

For Windows OS:

.venv\Scripts\activate

For Unix/MacOS Systems:

source .venv/bin/activate

1. Inject required libraries using uv: bash uv pip install -r requirements.txt

2. Generate a configuration file named .env containing your access token:

GOOGLE_API_KEY=your_api_key_here

Activating the Service Interface

Two primary pathways exist for initiating the MCP interface:

Method A: Direct Execution

To launch the MCP backend directly from the terminal:

bash uv run python server.py

Method B: Configuration for Claude.app Integration

For non-package-manager-dependent setup on Windows environments, embed the following structure into your Claude application settings:

"mcpServers": { "youtube": { "command": "C:\Path\To\Your\Project.venv\Scripts\python.exe", "args": ["C:\Path\To\Your\Project\server.py"], "env": { "GOOGLE_API_KEY": "your_api_key_here" } } }

When employing the uv toolchain on Windows, utilize this configuration:

"mcpServers": { "youtube": { "command": "uv", "args": ["--directory", "C:\Path\To\Your\Project", "run", "server.py"], "env": { "GOOGLE_API_KEY": "your_api_key_here" } } }

Exposed Toolset

The backend exposes the following discrete operational modules:

1. search-youtube: Executes a discovery operation on YouTube content based on supplied criteria.

2. Arguments:

   • query: The textual specification for the search string.
   • max_results: The upper bound on the quantity of returned items (default is set to 5).

3. get-transcript: Fetches the complete textual transcription for a specified video resource.

4. Arguments:

   • video_url: The direct Uniform Resource Locator pointing to the YouTube asset.

5. store-video-info: Persists the retrieved video metadata and its corresponding transcript into the designated vector repository.

6. Arguments:

   • video_url: The Uniform Resource Locator for the YouTube item.
   • metadata: Optional supplementary descriptive data pertaining to the video.

7. search-transcripts: Performs advanced conceptual matching against all previously cataloged transcript data.

8. Arguments:
   • query: The conceptual search phrase.
   • limit: The maximum enumeration of results to return (defaults to 3).

Utilization with MCP Consumers

This backend is interoperable with any software client adhering to the MCP specification, such as the Claude Desktop Application. The integrated tools are automatically recognized and made accessible to the client environment.

Illustrative Example

1. Initiate the service using one of the procedures detailed previously.
2. Launch the Claude Desktop Application.
3. Observe the wrench icon to confirm the availability of the YouTube utilities.
4. You may then issue directives such as:
5. "Locate video resources on Python programming tutorials"
6. "Acquire the textual record for this asset: [video_url]"
7. "Scan the saved video narratives for concepts related to neural network architecture"

Troubleshooting Guide

If operational anomalies arise:

1. Confirm that the Google API authorization credential is correctly positioned within the .env file.
2. Validate that all necessary software packages have been successfully integrated.
3. Ensure the backend service is actively executing and awaiting incoming network requests.
4. Scrutinize the service console output for any reported exception diagnostics.

Contribution Guidelines

We warmly invite external contributions! Please feel free to submit proposed enhancements via a Pull Request.

Licensing Information

This software is distributed under the terms of the MIT License; consult the LICENSE documentation for comprehensive details.