Welcome to GXtract’s documentation!¶
GXtract is a powerful and flexible Model Context Protocol (MCP) server with seamless GroundX tool integration. It enables efficient extraction of architectural and implementation details from various document types, with a focus on scientific papers and research articles.
This server is built with performance, scalability, and ease of use in mind, leveraging Python 3.12+, asyncio, and the FastMCP v2 library.
Key Features¶
High Performance: Asynchronous architecture using
asyncioand FastMCP v2.Cross-Platform: Runs on Windows (AMD64, ARM64) and Linux (AMD64, ARM64).
Extensible Tooling: Easily integrate new tools and functionalities.
GroundX Integration: Out-of-the-box tools for interacting with GroundX services (search, query, explain semantic objects).
Caching: Efficient caching mechanism for GroundX metadata to improve performance and reduce API calls.
Configurable: Flexible configuration via CLI arguments and environment variables.
Comprehensive Documentation: Built with Sphinx for clarity and ease of understanding.
Table of Contents¶
Getting Started
Tools
Development
- Development Guide for GXtract
- Prerequisites
- Project Structure {#project-structure}
- Setting up a Development Environment
- Coding Standards and Linting
- Running Tests (To Be Implemented)
- Building Documentation
- Creating a New Tool
- Versioning and Changelog
- Submitting Contributions (If Applicable)
- Key Modules
- Dependency Management
- Further Reading
- GXtract Architecture
- Changelog
API Reference
Project Info
Online Documentation¶
This documentation is also available online at https://sascharo.github.io/gxtract/.
The online documentation is automatically built and deployed whenever changes are pushed to the main branch of the repository, ensuring it’s always up to date with the latest features and improvements.