Model Context Protocol

SciRouter MCP Server

Give AI agents access to 25+ scientific computing tools. Claude, GPT, Cursor, and any MCP-compatible client can discover and call protein folding, molecular analysis, drug discovery, and vaccine design tools.

Quick Setup

Claude Desktop / Claude Code

Add this to your claude_desktop_config.json or .claude/settings.json:

{
  "mcpServers": {
    "scirouter": {
      "command": "python",
      "args": ["-m", "scirouter.mcp"],
      "env": {
        "SCIROUTER_API_KEY": "sk-sci-YOUR_KEY"
      }
    }
  }
}
json

Cursor

Add to .cursor/mcp.json in your project:

{
  "mcpServers": {
    "scirouter": {
      "command": "python",
      "args": ["-m", "scirouter.mcp"],
      "env": {
        "SCIROUTER_API_KEY": "sk-sci-YOUR_KEY"
      }
    }
  }
}
json

Python (direct API)

pip install scirouter
bash
from scirouter import SciRouter

client = SciRouter(api_key="sk-sci-YOUR_KEY")

# Fold a protein
result = client.proteins.fold("MVLSPADKTNVKAAWGKVGA...")
print(f"pLDDT: {result.mean_plddt}")

# Calculate molecular properties
props = client.chemistry.properties("CC(=O)Oc1ccccc1C(=O)O")
print(f"MW: {props.molecular_weight}")
python

Tool Directory38 tools

fold_protein5 cr

Predict 3D protein structure from amino acid sequence (ESMFold)

sequenceconfidence_threshold?
predict_protein_function5 cr

Predict protein function with GO terms and biological reasoning (BioReason-Pro)

sequenceorganism?include_reasoning?dna_sequence?
align_proteins1 cr

Pairwise sequence alignment (Needleman-Wunsch)

sequence_asequence_b
annotate_protein2 cr

Get UniProt annotations: domains, GO terms, drugs, diseases

uniprot_id
search_proteins3 cr

Ultrafast sequence search against SwissProt/UniRef50 (MMseqs2)

sequencedatabase?max_hits?e_value?
detect_pockets3 cr

Detect binding pockets in protein structures (fpocket)

pdb_stringn_pockets?
compare_structures2 cr

Superposition and RMSD/TM-score comparison of two structures

structure_astructure_b
molecular_properties1 cr

Calculate molecular properties from SMILES (RDKit)

smiles
convert_molecule1 cr

Convert between SMILES, InChI, MOL, InChIKey formats

inputfrom_formatto_format
molecule_similarity1 cr

Tanimoto fingerprint similarity between two molecules

smiles_asmiles_b

Example: Agent Conversation

Here's what it looks like when an AI agent uses SciRouter tools:

U
Analyze aspirin's drug properties and check if it would pass Lipinski's rules.
Tool call: molecular_properties(smiles="CC(=O)Oc1ccccc1C(=O)O")
Tool call: adme_predict(smiles="CC(=O)Oc1ccccc1C(=O)O")
Aspirin (MW: 180.2, LogP: 1.2, HBD: 1, HBA: 4) passes all Lipinski rules with 0 violations. ADMET screening shows good oral absorption (HIA: positive), no BBB penetration, and low hERG risk. It's classified as drug-like with no PAINS alerts.

Ready to give your AI agent superpowers?

Get a free API key with 500 credits/month. No credit card required.