Github Agentic Chat MCP

🚀 GitHub Agentic Chat MCP Server

This is a GitHub agentic chat MCP (Model Context Protocol) server implemented in Go. It provides tools for interacting with GitHub via natural language and integrates vector search functionality.

🚀 Quick Start

This server allows users to interact with GitHub using natural language and perform vector searches. Follow the installation and configuration steps below to get started.

✨ Features

• Search GitHub repositories
• Create issues
• Vector search functionality
  • Add documents to the vector store
  • Perform semantic searches on stored documents
• Scalable architecture for easy addition of more features

📦 Installation

Prerequisites

• Go 1.21 or higher
• PostgreSQL database (with the pgvector extension installed)
• GitHub personal access token
• OpenAI API key
• Claude Desktop or other MCP-compatible clients

Steps

1. Clone the repository:

git clone https://github.com/akhidasTech/github-agentic-chat-mcp.git
cd github-agentic-chat-mcp

2. Set environment variables:

export GITHUB_TOKEN=your_github_token_here
export DATABASE_URL=postgres://user:password@localhost:5432/dbname
export OPENAI_API_KEY=your_openai_api_key_here

3. Configure the PostgreSQL database and install the pgvector extension:

CREATE EXTENSION vector;

4. Build the server:

go build -o bin/github-agentic-chat-mcp

5. Configure Claude Desktop: Edit the ~/Library/Application Support/Claude/claude_desktop_config.json file and add the following content:

{
    "mcpServers": {
        "github-chat": {
            "command": "/absolute/path/to/bin/github-agentic-chat-mcp"
        }
    }
}

6. Restart the Claude Desktop client.

💻 Usage Examples

GitHub Tools

search_repositories

Search GitHub repositories using a query string.

create_issue

Create a new issue in a specified GitHub repository.

Vector Search Tools

add_to_vector_store

Add a document to the vector store for semantic search support.

Parameters:

• content: The text content to be stored.
• metadata: A JSON string containing content metadata.

Example:

{
    "content": "This is a document about GitHub Actions",
    "metadata": "{\"type\": \"documentation\", \"tags\": [\"github\", \"ci-cd\"]}"
}

vector_search

Perform a semantic search on stored documents.

Parameters:

• query: The search query text.
• limit: The maximum number of results to return (default is 5).

🤝 Contribution

Contributions are welcome! Feel free to submit code changes via Pull Requests.

📄 License

MIT License