Molecule analysis
Canonicalization, descriptors (MW, logP, QED, TPSA), scaffold diversity, and novelty audits — computed with RDKit.
SciAgentKit gives Claude, Cursor, Gemini and Codex deterministic tools for drug discovery — descriptors, docking, molecular dynamics, and reports — instead of letting them invent numbers.
Deterministic, reproducible, manifest-backed. Each tool runs real science and records exactly how.
Canonicalization, descriptors (MW, logP, QED, TPSA), scaffold diversity, and novelty audits — computed with RDKit.
UniProt resolution, PDB ranking with mutation-aware alignment, coverage and resolution scoring.
Pocket detection (crystal-ligand, fpocket, P2Rank) and Vina/GNINA/smina docking with ranked poses.
OpenMM job setup, solvation and ions, plus trajectory RMSD/RMSF and ProLIF interaction analysis.
Professional Markdown / PDF / DOCX reports with tables, figures, and a reproducibility manifest.
Every run hashes its inputs and pins tool versions. Licensed, offline-verifiable, audit-ready.
Skills decide the workflow. MCP runs the science. Reports preserve the evidence.
Literature + UniProt/PDB ranking.
Ligand 3D + receptor cleanup.
Solubility, drug-likeness, PAINS/Brenk alerts.
Pocket detection + scoring.
OpenMM MD + trajectory.
Target-aware: toxicity + BBB weighted.
Assay + cell-line recommendations.
Signed PDF with everything.
Start free for 3 days. Upgrade any time — your license updates instantly.
Try the workflow, no card required.
For a single researcher.
One flat price for your whole lab (up to 5 seats).
Team is one flat price for up to 5 seats. Add Expert review at checkout — a specialist evaluates your results (charged per member for Team).
Install, add your API key, and connect it to your AI agent.
Start a free 3-day trial and copy your key from My account.
Full pipeline (all engines) with conda:
curl -O https://sciagentkit.com/environment.yml conda env create -f environment.yml conda activate sciagentkit export SCIAGENTKIT_API_KEY=sak_your_key
Analysis only: pip install "sciagentkit[mcp]"
claude mcp add sciagentkit -s user \ --env SCIAGENTKIT_API_KEY=sak_your_key \ -- python -m sciagentkit.servers.rdkit_server
sciagent init-agent cursor # restart Cursor from the project folder
sciagent init-agent gemini # in Gemini CLI: /mcp /skills
sciagent init-agent codex # in Codex: /mcp /skills (or $smiles-analysis ...)
Paste this into Claude Code, Codex, or Gemini CLI — it installs everything and connects itself (replace the key first):
Set up SciAgentKit for me. Steps: 1) Create a conda env "sciagentkit" with python 3.12 and install from conda-forge: rdkit openbabel autodock-vina fpocket openmm pdbfixer mdanalysis prolif 2) Then run: pip install "sciagentkit[mcp]" 3) Set the environment variable SCIAGENTKIT_API_KEY=sak_your_key 4) Register the MCP server: claude mcp add sciagentkit -s user --env SCIAGENTKIT_API_KEY=sak_your_key -- python -m sciagentkit.servers.rdkit_server 5) Run "sciagent doctor" and "sciagent license-verify" and show me the output.
Built by computational drug-discovery researchers.
Founder & Developer
Computational drug-discovery researcher at Bahçeşehir University. Designs and builds the SciAgentKit toolkit and its MCP integration.
Founder & Scientific Advisor
Professor of computational drug design and molecular modeling; Dean of the School of Pharmacy; founder of Istanbul MedChem.
Spin up a license in seconds and wire it into Claude, Cursor, Gemini or Codex.