The Medical Calculator is a DDD-architected service designed to address the challenges AI agents face when performing medical calculations, such as hallucination risks, version confusion, and the lack of discovery mechanisms. It offers 82 peer-reviewed, validated clinical scoring tools, ensuring accuracy by citing original research papers. The project employs a human-in-the-loop, AI-augmented workflow that involves human expert domain specification, AI-driven guideline search, systematic extraction of recommended scoring systems, primary source validation of formulas, and precise implementation. This service enables AI agents to accurately search by specialty, condition, or clinical question and perform calculations via a standardized protocol, making it a reliable backend for intelligent medical applications.
Key Features
01Clean DDD Architecture using Onion architecture for clear separation of concerns
02Type Safe implementation with full Python type hints and dataclass entities
03MCP Native Integration for seamless AI agent communication
04Evidence-Based formulas with citations to original peer-reviewed research
05Intelligent Tool Discovery with a two-level key system for smart tool selection
060 GitHub stars
Use Cases
01Integrating accurate medical calculations into AI agents and LLMs
02Developing custom medical web applications and scripts requiring validated clinical tools
03Enhancing AI copilots (e.g., VS Code Copilot, Claude Desktop) with a reliable source of medical calculators
