home / mcp / biomcp mcp server

BioMCP MCP Server

BioMCP: Biomedical Model Context Protocol

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "genomoncology-biomcp": {
      "command": "uv",
      "args": [
        "run",
        "--with",
        "biomcp-python",
        "biomcp",
        "run"
      ],
      "env": {
        "CBIO_TOKEN": "YOUR_API_TOKEN",
        "ONCOKB_TOKEN": "YOUR_ONCOKB_TOKEN",
        "CBIO_BASE_URL": "https://www.cbioportal.org/api",
        "BIOMCP_METRICS_ENABLED": "false",
        "BIOMCP_USE_CONNECTION_POOL": "true"
      }
    }
  }
}

BioMCP is an open source MCP server that connects AI assistants to authoritative biomedical data sources, enabling precise, context-rich queries about clinical trials, literature, and genomic variants. It provides structured data access and a unified interface so you can build biomedical reasoning into your AI workflows with confidence and reproducibility.

How to use

To use BioMCP with an MCP client, run the BioMCP server in your chosen transport mode and point your client to the server using the provided endpoint or stdio interface. Start by launching the server in the mode your environment supports, then issue queries through your MCP client to retrieve articles, trials, variants, genes, diseases, or drugs. BioMCP exposes tools that perform searches, fetch details, and retrieve structured biomedical data across multiple sources, and it automatically augments searches with domain data such as cBioPortal summaries when applicable.

Operational workflow

1) Start the server in your preferred mode (stdio for local development or http for network access). 2) Choose a tool or endpoint to perform a task, such as searching for literature on a gene, retrieving trial details, or fetching a gene’s annotations. 3) Combine domain-specific queries to refine results, then fetch full records when needed. 4) Use the results to inform biomedical analyses, experiments, or decision support in your AI workflows.

Available tools

think

A guided, multi-step thinking workflow that breaks complex biomedical questions into smaller parts and tracks progress across thoughts.

search

Unified and domain-based search across genes, variants, diseases, trials, and articles with optional schema and explanation features.

fetch

Retrieve full details for a single article, trial, or variant with domain-specific options for depth and sections.

article_searcher

Search PubMed/PubTator3 and preprint sources for biomedical literature.

article_getter

Fetch detailed article information using PMID or DOI.

trial_searcher

Search ClinicalTrials.gov or NCI CTS API with optional sources and advanced filters.

trial_getter

Fetch complete trial details, including protocol, locations, outcomes, and references.

trial_protocol_getter

Fetch only the protocol information for a trial.

trial_references_getter

Fetch publications associated with a trial.

trial_outcomes_getter

Fetch outcome measures and results for a trial.

trial_locations_getter

Fetch trial site locations and contacts.

variant_searcher

Search MyVariant.info for genetic variant annotations.

variant_getter

Fetch comprehensive variant details across annotations.

nci_organization_searcher

Search NCI organization database.

nci_organization_getter

Get NCI organization details by ID.

nci_intervention_searcher

Search NCI interventions (drugs, devices, procedures).

nci_intervention_getter

Get intervention details by ID.

nci_biomarker_searcher

Search biomarkers used in trial eligibility criteria.

nci_disease_searcher

Search NCI disease vocabulary.

gene_getter

Fetch real-time gene information from MyGene.info.

disease_getter

Get disease definitions and synonyms from MyDisease.info.

drug_getter

Get drug/chemical information from MyChem.info.