QuantaMCP | One API Key. Every Molecule, Protein, and Gene.
Setup Tools QBench Enterprise Helios Dynamics Request Access

Cloud Platform

One API key. Every molecule, protein, and gene.

QuantaMCP speaks the Model Context Protocol. Connect once, and any MCP-compatible AI — Claude, ChatGPT, Cursor, Copilot — gains access to 122M+ compounds, 214M+ predicted structures, and 30+ life sciences databases.

Request Access View setup
122M+ Enriched compounds
<50ms Response time
27+ MCP tools
mcp_config.json 
{
  "mcpServers": {
    "quantamcp": {
      "type": "streamable-http",
      "url": "https://quantamcp.com/quantamcp/mcp",
      "auth": "oauth2"
    }
  }
}
// That's it. 27+ tools, zero SDK.

How QuantaMCP thinks

From natural language to structured intelligence in three stages.

01 / Ingest

Route to the right tool

Your question arrives as natural language. The MCP protocol parses intent and routes it to the right tool — predict_admet, search_similar, check_compliance — without you specifying endpoints or parameters.

02 / Compute

Parallel intelligence

Pre-computed predictions across 122M compounds return instantly. Novel compounds trigger real-time ML inference. Regulatory screening runs in parallel across 8 jurisdictions. Literature and patent search executes concurrently.

03 / Return

Structured results in <50ms

ADMET profiles, compliance flags, literature citations, structural alerts — returned as structured data your AI can reason over. Not a wall of text. Not a dashboard to click through. Intelligence, inline.

API reference

Every tool available through the MCP protocol. No SDK required.

27 tools / 5 resources / 4 prompts — included with every account

Molecular Intelligence
get_molecule_profile
Full profile with ADMET + compliance. Pre-computed for 122M+ compounds.
get_molecule_info
Basic RDKit properties: MW, LogP, TPSA, HBD/HBA.
calculate_properties
On-demand: MW, LogP, LogD, TPSA, SA Score, Lipinski/Veber.
get_3d_properties
32+ 3D properties from conformer generation.
search_similar
Tanimoto similarity search across 122M+ compounds.
filter_molecules
Filter by MW, LogP, QED. Exclude controlled substances.
batch_profile
Profile up to 100 compounds per call.
predict_admet
40+ ML predictions: cardiotoxicity, CYP450, hepatotoxicity.
optimize_molecule
NVIDIA MolMIM variants. Target QED, LogP. Auto-FAVES.
Regulatory & Compliance
check_compliance
FAVES: DEA, FDA, EPA, CWC, EU REACH. 8 jurisdictions.
screen_library
Batch FAVES screening. Up to 1,000 molecules per call.
Research & Literature
search_literature
14,000+ curated drug discovery papers.
search_patents
2,400+ USPTO pharmaceutical patents.
search_biorxiv
250K+ bioRxiv/medRxiv preprints.
search_chembl
ChEMBL: 2.4M compounds with bioactivity data.
search_clinical_trials
500K+ trials from ClinicalTrials.gov.
Structure Prediction
get_protein_structure
Smart resolver: names, PDB IDs, or sequences → structures.
predict_structure
OpenFold3 prediction with confidence scores.
get_structure_result
Retrieve completed predictions. Download PDB files.
Structural Biology Coming Soon
query_uniprot
Protein annotations, domains, and cross-references.
query_pdb
Experimental structures. Filter by resolution, method.
query_alphafold
214M+ predicted structures with per-residue confidence.
Genomics Coming Soon
query_refseq
Gene sequences and coordinates from RefSeq/GenBank.
query_expression In Development
GEO/Expression Atlas: differential expression data.
Data Connectors & Platform
pull_from_source
Databricks/Snowflake → ADMET + FAVES → push back.
push_to_destination
Write to Snowflake, BigQuery, Databricks, Supabase.
get_platform_info
Database stats, capabilities, platform version.
get_credit_usage
Credit balance and usage statistics.

See it work

A real query, real tools, real results. This is what QuantaMCP returns when you ask a question.

Claude + QuantaMCP
You
What's the safety profile of compound CC(=O)Oc1ccccc1C(=O)O?
QuantaMCP
Tools: get_molecule_profile predict_admet check_compliance search_literature

Aspirin (acetylsalicylic acid) — safety profile:

QED0.55
FAVES8/8 passed
HepatotoxLow (0.12)
hERGLow (0.08)
CYP2D6Negative
Lipinski0 violations
Patrono et al. "Antiplatelet Drugs" (2017)
Vane. "Prostaglandin Synthesis Inhibition" (1971)
Rothwell et al. "Aspirin and Cancer Risk" (2012)

Works with your AI

Claude ChatGPT Cursor Windsurf GitHub Copilot VS Code Claude Code Any MCP client

Built for developers

SDKs to install
Endpoints to memorize
Rate limits to fight
Dashboards to learn
Context switching

One URL, one API key, and your AI assistant becomes a molecular intelligence platform. New tools ship server-side. Your setup never goes stale.

Connect in 60 seconds

Works with any MCP-compatible AI client. Pick your platform and paste the URL.

Server URL: https://www.quantamcp.com/quantamcp/mcp Transport: Streamable HTTP Auth: OAuth 2.0 (automatic) # Claude Desktop/Web — Settings → Connectors → Add → paste URL # Claude Code — claude mcp add --transport http quantamcp [URL] # Cursor/Windsurf — Settings → MCP → Add server → paste URL # Any MCP client — add the URL, authenticate, done

Then ask your AI: "What's the ADMET profile of aspirin?"

Start building with QuantaMCP

Request API access and connect your AI assistant in under a minute.

Request Access