Every molecule, already understood
Toxicity. Drug-likeness. Regulatory status. Drug interactions. Pre-computed across 122 million compounds so your team never waits. Set a direction — the platform identifies targets, optimizes leads, docks candidates, and stratifies patients — without asking permission at every step.
NovoMCP
Your AI already knows chemistry. 122 million compounds with ADMET, compliance, and drug-likeness pre-computed — accessible from Claude, ChatGPT, Cursor, or any MCP-compatible assistant. Connect once. The platform handles the rest.
- 31 tools — zero configuration
- Target-to-clinic autonomous pipelines
- GPU docking and molecular dynamics
- Compliance inline across 8 jurisdictions
- Snowflake, Databricks, BigQuery connectors
NovoWorkbench
The molecular workstation that lives on your machine. 3D visualization, local RDKit compute, and FAVES compliance screening — online or offline. Connect NovoMCP when you need the full platform.
- 3D visualization without browser limits
- Local compute — no Python required
- FAVES compliance, online and offline
- Multi-provider AI (Claude, GPT-5, Ollama)
- Optional NovoMCP cloud integration
What your team stops doing
The work that used to take days already has answers waiting.
Waiting for ADMET results
Every molecule in the index already has its absorption, metabolism, and toxicity profile computed. 84 properties across 122 million compounds. Your team asks — the answer is already there. Without model queues, compute budgets, or cold starts.
ADMET Radar — Aspirin
CC(=O)Oc1ccccc1C(=O)O
85%
Abs
72%
Dis
91%
Met
78%
Exc
64%
Tox
Running compliance as a separate step
DEA, FDA, EPA, CWC, EU REACH, BTWC, Australia, OPCW — eight jurisdictions screened as results are returned, not in a meeting two weeks later. Controlled substance detection, structural alerts, PAINS filtering. Compliance runs inside the pipeline — without a manual checkpoint at the end.
FAVES Compliance — Aspirin
CC(=O)Oc1ccccc1C(=O)O
DEA
CLEARFDA
CLEARCWC
CLEAREPA
CLEAREU REACH
CLEAROverall Status
CLEARPAINS: 0
Structural Alerts: 0
Whitelisted FDA-approved drug
Searching five databases separately
14,000 curated papers. 2,400 patents. 250,000 preprints. 500,000 clinical trials. 2.4 million bioactive compounds from ChEMBL. One query gives your team the complete landscape — without five logins, five search syntaxes, five result formats.
Unified Research — Aspirin
5 sources · 1 queryAspirin and cardiovascular disease prevention: a systematic review
Lancet · 2024 · Cited 142
Novel aspirin formulations with improved bioavailability
USPTO · US2023/0142891 · Filed 2023
ADAPTABLE: Aspirin Dosing in Cardiovascular Disease
Phase IV · Completed · n=15,076
COX-2 selectivity and platelet aggregation: mechanistic insights
bioRxiv · 2024 · doi:10.1101/2024.03.15
Bioactivity: COX-1 IC₅₀ = 1.67 µM, COX-2 IC₅₀ = 278 µM
CHEMBL25 · 1,247 assays · 892 targets
Handing off between optimization, docking, and simulation
Scaffold hopping generates structurally novel variants. GPU-accelerated docking scores them against your protein target. Molecular dynamics validates binding stability. Every candidate compliance-screened before it reaches your team — without context switching between tools.
Building enrichment pipelines from scratch
Pull compound libraries from Snowflake, Databricks, BigQuery, or Supabase. Run ADMET and compliance enrichment. Push results back. Your existing infrastructure gains molecular intelligence — without new ETL code, without new dashboards, without a data engineering sprint.
One sentence. The platform does the rest.
Lead Optimization
"Optimize this kinase inhibitor for oral bioavailability." Scaffold hopping, ADMET prediction, GPU docking, and compliance screening — orchestrated from one sentence.
Safety Assessment
"Is this compound safe to develop?" Eight regulatory jurisdictions and structural alert databases checked before you finish reading the response.
Literature Review
"What's known about CDK4/6 inhibitors in breast cancer?" Papers, patents, preprints, trials, and bioactivity — one query instead of five separate searches.
Virtual Screening
"Find molecules similar to osimertinib with QED above 0.6 and no PAINS." 122 million compounds filtered by structure, properties, and compliance in seconds.
Structure Prediction
"Predict the structure of this kinase domain." Protein folding with per-residue confidence scores — without managing compute or navigating protein databases.
Data Pipeline
"Pull my compound library from Databricks. Run ADMET and FAVES. Push results back." Your warehouse compounds, molecularly enriched. One prompt.
Discovery Funnel
"Find targets for glioblastoma and run the full discovery funnel." Target identification, lead optimization, GPU docking, MD simulation, patient stratification — autonomously, from one sentence.
How it works
Connect in sixty seconds
One URL, one API key. Works with Claude, ChatGPT, Cursor, Windsurf, and any MCP-compatible assistant. No SDKs to install, no dashboards to learn.
Describe what you need
Say it in plain language. Your AI selects the right tools, chains them together, and pursues the objective — without asking permission at every step.
Get answers with compliance built in
Every response screened against eight regulatory jurisdictions. Lead optimization produces only compliant variants. The pipeline has guardrails at every step — not just the last one.
The quiet engine behind thoughtful discovery
One URL. One API key. Your AI already knows chemistry.