Medical Research Toolkit

Query 14+ biomedical databases for drug repurposing, target discovery, clinical evidence, and literature research — all via a unified MCP endpoint.

⚡ 30-Second Start

CODEBLOCK0

That's it! You now have approved and investigational drugs for the disease.

Quick Recipes

Find Drugs for a Disease

CODEBLOCK1

Returns: Approved + investigational drugs with max phase reached

Find Disease Targets

CODEBLOCK2

Returns: Top disease targets ranked by evidence strength (0-1 score)

Search Literature

CODEBLOCK3

Returns: PubMed articles on myasthenia gravis immunotherapy

Find Active Clinical Trials

CODEBLOCK4

Returns: Actively recruiting trials for the disease

Check Drug Safety

CODEBLOCK5

Returns: FDA adverse event reports—check for serious outcomes, death, hospitalization

What You Get

14+ Integrated Databases

Use Cases

🧬 Drug Repurposing

1. 1. Find disease targets (OpenTargets)
2. Search for drugs targeting those genes (ChEMBL)
3. Check ongoing trials (ClinicalTrials)
4. Verify safety (OpenFDA, PubMed)

🔬 Target Discovery

1. 1. Find disease associations (OpenTargets, GWAS)
2. Get pathway context (Reactome, KEGG)
3. Review literature (PubMed)
4. Check protein properties (UniProt)

📋 Clinical Evidence Review

1. 1. Search trials (ClinicalTrials.gov)
2. Find literature (PubMed)
3. Check FDA data (OpenFDA)

📊 Literature Mining

1. 1. PubMed: 35M+ articles searchable by gene, disease, drug, chemical
2. Preprints: bioRxiv, medRxiv
3. Filter by keywords, date, study type

API Endpoint

Production (No setup needed):
CODEBLOCK6

All 14+ databases unified into one endpoint.

Running Locally:

pip install medical-mcps
medical-mcps
# Available at: http://localhost:8000/tools/unified/mcp

Complete References

See detailed guides for each database:

• - PubMed Guide — Literature search (genes, diseases, keywords)
• ClinicalTrials Guide — Find active trials
• ChEMBL Guide — Drug-target data & bioactivity
• OpenTargets Guide — Disease-target associations
• OpenFDA Guide — Drug safety & adverse events
• OMIM Guide — Genetic diseases (requires API key)
• Other APIs — Reactome, UniProt, KEGG, GWAS, etc.

Workflow Example

Complete Drug Repurposing Pipeline:

See drug-repurposing-workflow.md for step-by-step example with all 8 steps + curl commands.

API Keys

Most APIs are free, no key required. Optional keys for higher rate limits:

Rate Limits & Caching

• - No authentication for production endpoint (public)
• Rate limits: Generous (~1000+ requests/day per database)
• Caching: Automatic 30-day HTTP caching (RFC 9111)
• Cost: $0 (all databases public or researcher-accessible)

Common Patterns

Batch Query Loop

CODEBLOCK8

ID Conversion

Need to convert IDs between databases?

CODEBLOCK9

Returns: Equivalent IDs in NCBI Gene, Uniprot, Ensembl, MyGene, etc.

Troubleshooting

No results?

• - Try alternative terms (gene symbol vs protein name)
• Check spelling
• Use standardized IDs (gene symbols, disease names)
• Some APIs use specific ID formats (EFO vs MONDO, etc.)

Too many results?

• - Add filters (max_results, phase, recruitment status)
• Use limit or size parameters
• Combine with other databases to narrow focus

API key errors?

• - OMIM requires API key — get from https://omim.org/api
• Other databases optional — request key if hitting rate limits

Next Steps

1. 1. Pick a use case (drug repurposing, target discovery, etc.)
2. Read the relevant database guide from References section
3. Copy a quick recipe from above
4. Customize parameters for your disease/gene/drug
5. Scale up — build your full research workflow

Resources

• - Source Code: https://github.com/pascalwhoop/medical-mcps
• Author: Pascal Brockmeyer (@pascalwhoop)
• Organization: Every Cure (https://www.everycure.org)
• License: MIT

Getting Help

• - Database not working? → Check Troubleshooting
• Want detailed guide? → See Complete References
• Need a workflow example? → See drug-repurposing-workflow.md
• Question about OpenClaw? → See OPENCLAW-USAGE.md