We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/isc-tdyar/medical-graphrag-assistant'
If you have feedback or need assistance with the MCP directory API, please join our Discord server
# Medical GraphRAG Assistant Constitution
## Core Principles
### I. IRIS-Native
The system MUST prioritize InterSystems IRIS native features (SQL, Vector Search, Graph traversals) over external logic. Database logic MUST be parameterized to prevent injection. The project MUST use `intersystems-irispython` as the official Python driver (not `iris` or `intersystems-iris`).
### II. Agent-Centric Design
All features MUST be exposed via the Model Context Protocol (MCP) as tools. Tools MUST return structured JSON to ensure predictable LLM consumption.
### III. Medical Data Integrity
The system MUST respect FHIR R4 resource structures. Clinical data retrieved via tools MUST be verifiable against the source IRIS database.
### IV. Observability & Memory
Agent corrections and preferences MUST be persisted in the semantic memory system. Every search execution MUST log its retrieval strategy (e.g., RRF scores) for transparency.
### V. Browser-First Verification
UX features MUST be verified via browser automation (Playwright) to ensure the Streamlit frontend correctly renders complex reactive visualizations.
## Governance
All implementation tasks must verify compliance with these principles. Complexity must be justified in the feature plan.
**Version**: 1.0.0 | **Ratified**: 2026-01-02