Agent2Agent MCP Server

Example Use Cases

• 1

  Generate SQL from clinical natural-language prompts against an OMOP CDM while incorporating specific medical concept codes (conditions, drugs, measurements).

• 2

  Validate generated SQL against OMOP CDM validation rules before execution to reduce unsafe or invalid queries.

• 3

  Execute validated SQL on a PostgreSQL OMOP instance and return results (CSV) and natural-language explanations or answers derived from query results.

Description

Agent2Agent MCP Server provides tools for interacting with OMOP CDM databases: generating SQL from natural language (with integrated medical concept codes), validating queries against OMOP rules, executing queries and returning CSV results, and producing natural-language explanations and answers. It integrates an LLM (via Ollama) for SQL generation and explanation while using FastMCP as the protocol layer. The server is intended to help automate clinical query generation and validation workflows and to support agents that need to reason about clinical data and translate intent into safe, validated SQL.