Hackage Documentation MCP

Overview

This Machine-Controlled Program (MCP) provides an interface for AI assistants to access Haskell documentation from Hackage. It allows AI models to retrieve precise, up-to-date documentation for specific Haskell modules, improving the AI's ability to assist with Haskell programming despite limited training data in this domain.

Problem Statement

AI language models often have insufficient training data on Haskell compared to more mainstream programming languages. This results in:

• Less accurate code generation for Haskell
• Difficulty understanding Haskell's unique features and abstractions
• Inaccurate or outdated information about Haskell libraries and modules

This tool bridges that gap by enabling AI to retrieve authoritative documentation from Hackage in real-time.

Features

• Retrieve documentation for specific modules from any package on Hackage
• Specify exact package versions to ensure accuracy
• Clean, text-based output that's easily consumable by AI assistants

Installation

1. Ensure you have Python 3.7+ installed
2. Install the required dependencies:

   pip install httpx beautifulsoup4 fastmcp
   

Usage

You can use this MCP with any AI assistant that supports the MCP protocol.

Running the Server

python main.py

This will start the MCP server using stdio transport.

API

The MCP exposes the following tool:

get_docs(package, version, module)

Retrieves documentation for a specific Haskell module.

Parameters:

• package (string): The Hackage package name (e.g., "lens", "yesod-persistent")
• version (string): The package version (e.g., "5.0.0", "2.10.0")
• module (string): The module name to look up (e.g., "Control.Lens", "Database.Persist")

Returns:

• The extracted documentation text for the specified module, or an error message if retrieval fails

Examples:

# Get documentation for Control.Lens from lens 5.0.0
get_docs("lens", "5.0.0", "Control.Lens")

# Get documentation for Data.Vector.Algorithms from vector-algorithms 0.8.0
get_docs("vector-algorithms", "0.8.0", "Data.Vector.Algorithms")

Use Cases

1. Learning Haskell libraries: Get detailed information about specific modules while working with the AI
2. Understanding type signatures: Access accurate type information for functions in Haskell modules
3. Exploring module hierarchies: Understand how modules are organized within Haskell packages
4. Verifying AI-generated code: Compare AI suggestions with official documentation

Benefits

• Provides AI with access to accurate, up-to-date Haskell documentation
• Improves the quality of AI-generated Haskell code
• Enhances the AI's ability to explain Haskell concepts with authoritative references
• Reduces hallucinations when discussing Haskell libraries

Limitations

• Requires an internet connection to access Hackage
• Only provides documentation content, not the implementation code
• Documentation quality varies between Haskell packages

Contributing

Contributions to improve this tool are welcome! Some potential areas for enhancement:

• Adding support for downloading and parsing package source code
• Implementing caching to reduce network requests
• Expanding to support other Haskell documentation sources

License

[Specify your license here]