VecMed-MCP: Milvus Vector Database for Medical Data

Overview

VecMed-MCP is a lightweight framework for building and maintaining a Milvus-based vector database tailored for medical and rare disease research.
It provides end-to-end tools for database initialization, data ingestion, semantic search, LLM-driven summarization, and automated updates, enabling researchers to efficiently manage large-scale biomedical literature.

Key Features

MilvusDB Integration – Ready-to-use Docker setup for fast deployment

Rare Disease Dataset – Includes over 160,000 PubMed records preloaded

Semantic Search – Vectorized search across medical abstracts and metadata

LLM Summarization – Automatic summarization of search results for quick insights

Scheduled Updates – Maintain an up-to-date database with cron-based automation

WebUI (Attu) – Visual interface for browsing collections and schema management

Workflow

Initialize MilvusDB with provided schema setup scripts

Ingest biomedical data (e.g., PubMed articles) into the vector store

Query and search through vector embeddings for rare disease insights

Summarize results with integrated LLM pipelines

Automate updates for continuous knowledge enrichment

Applications

Rare disease literature search and knowledge retrieval

Clinical decision support with medical text embeddings

Integration into MCP-compatible agents, workflows, or LLM-powered systems