genomoncology/biomcp
BioMCP: Biomedical Model Context Protocol
What it solves
BioMCP provides a unified interface to access and query approximately 30 trusted biomedical sources (such as PubMed, ClinVar, and OncoKB) without needing to navigate multiple different APIs, identifiers, and search habits. It simplifies the process of gathering evidence across genes, variants, drugs, diseases, and proteins for researchers and clinicians.
How it works
It operates as a single CLI binary with a consistent command grammar. It can be used as a standalone tool or as an MCP (Model Context Protocol) server, allowing AI agents like Claude Code and Codex to access these biomedical tools directly. The system supports "pivoting" between related entities (e.g., moving from a gene to its associated trials) and provides focused output through selectable sections.
Who it’s for
Biomedical researchers, clinicians, and AI agent developers who need a streamlined way to retrieve evidence-oriented results from public biomedical data and perform local study analytics on cBioPortal-style datasets.
Highlights
- Unified Access: One grammar to reach ~30 sources including PubMed, ClinVar, and ClinicalTrials.gov.
- MCP Server Integration: Native support for AI agents via the Model Context Protocol.
- Cross-Entity Pivoting: Quickly move between genes, variants, drugs, and diseases without rebuilding filters.
- Local Study Analytics: Built-in tools for querying cohorts, survival, and co-occurrence in local datasets with SVG/PNG chart generation.
- Evidence Mapping: Tools to follow paper trails via citations, references, and recommendations.