BioContextAI Knowledgebase MCP
A single MCP server that fronts 14+ biomedical APIs as a unified retrieval layer — antibody and protein identifier resolution, pathway and interaction lookups, predicted structures, drug and clinical trial cross-references.
| Type | MCP server |
| Supplier | BioContextAI (Heidelberg-led OSS initiative) |
| Availability | GA — v0.2.1 (May 2026) |
| Pricing | Free / OSS (Apache-2.0). Underlying data sources keep their own licenses; KEGG restricts commercial remote hosting. |
| Capabilities | Read-only |
| Verified | works · 2026-07-20 |
| Security | cleared · 2026-07-20 — provenance matches biocontext-ai, PyPI biocontext_kb 0.2.1, Apache-2.0, no advisories |
The server is published as the biocontext_kb package and is most easily run with uv (which handles its own Python). To verify it starts:
uvx biocontext_kb@latest
Then register it as an MCP server (Claude Code launches the process itself — don’t keep the verification command running):
- Claude Code (stdio, recommended):
claude mcp add --transport stdio biocontext_kb -- uvx biocontext_kb@latest - Claude Desktop (stdio) — add to
claude_desktop_config.json:{ "mcpServers": { "biocontext_kb": { "command": "uvx", "args": ["biocontext_kb@latest"], "env": { "UV_PYTHON": "3.12" } } } } - Docker (HTTP) — no image is published to a registry, so you must build it from a clone first. The container is a long-lived service: leave it running in its own terminal.
git clone https://github.com/biocontext-ai/knowledgebase-mcp cd knowledgebase-mcp docker build -t biocontext_kb:latest . docker run -p 127.0.0.1:8000:8000 biocontext_kb:latestThen, in a second terminal:
claude mcp add --transport http biocontext_kb http://127.0.0.1:8000/mcp - Remote test endpoint (no install):
claude mcp add --transport http biocontext_kb https://biocontext-kb.fastmcp.app/mcp
What it does
Unified access to:
- Antibody / immunology: Antibody Registry, PRIDE
- Protein / structure: UniProt, InterPro, AlphaFold DB, Human Protein Atlas
- Interactions / pathways: STRING, Reactome, KEGG
- Targets / drugs: Open Targets, Drugs@openFDA
- Single-cell context: PanglaoDB
- Genomes / literature: Ensembl, EuropePMC, bioRxiv
- Ontologies: OLS
- Trials: ClinicalTrials.gov
Primary use cases: Antibody and reagent identifier resolution; immune protein and pathway context; pulling AlphaFold structures for antigen-binding analysis; building integrated knowledge graphs for a target list.
Notes
stdio (dev) and HTTP/uvicorn (prod) transports. No auth for the server itself, but it respects upstream rate limits. IEDB is not yet integrated.
The hosted biocontext-kb.fastmcp.app endpoint is published by the project for testing only and is subject to fair use — self-host for anything you depend on.
Sources
Installed this tool?
Share feedback — install path, OS, errors, workarounds. The form opens with this tool pre-selected and a link back to this page.