Hanzo

🚀 Hanzo MCP Project Guide

This guide provides a comprehensive introduction to the Hanzo MCP project, including installation, configuration, quick start, and feature overview.

🚀 Quick Start

Initialize the MCP Service

import asyncio
from hanzo_mcp.mcp_server import HanzoMCP

async def main():
    # Create a Hanzo MCP instance
    mcp = HanzoMCP(
        allowed_paths=["/path/to/project"],
        embedding_provider="voyage"  # Supported: 'openai', 'anthropic', 'sentence_transformer'
    )

    # Start the service
    await mcp.start()

# Run the main function
if __name__ == "__main__":
    asyncio.run(main())

✨ Features

Basic Operations

1. File Indexing

Semantically index files in a specified directory:

result = await dev(
    ctx,
    operation="vector_index",
    path="/path/to/project",
    recursive=True,
    file_pattern="*.py"  # Optional: Index only Python files
)

2. Semantic Search

Query the codebase using natural language:

result = await dev(
    ctx,
    operation="vector_search",

📦 Installation

Install Dependencies

To use Hanzo MCP, first install the necessary Python packages:

pip install hanzo-mcp

If you need additional features (such as vector storage), add the following dependencies as needed:

• VoyageAI Support:

  pip install voyageai
  

• SentenceTransformer Support:

  pip install hanzo-mcp[vector,sentencetransformer]
  

Configure Environment Variables

Set the corresponding environment variables based on the selected embedding provider:

• VoyageAI:
  • Configuration variables: VOYAGE_API_KEY or CHROMA_VOYAGE_API_KEY
• OpenAI:
  • Configuration variables: OPENAI_API_KEY or CHROMA_OPENAI_API_KEY
• Anthropic:
  • Configuration variables: ANTHROPIC_API_KEY or CHROMA_ANTHROPIC_API_KEY