Screen More
Compounds.
Spend Less on Wet Lab.
AI-powered in silico toxicology for drug discovery teams. Submit a SMILES string, get 50+ regulatory-grade ADMET predictions in minutes — OECD-compliant, with QMRF documentation ready for regulatory submission.
Wet-lab screening is slow and expensive.
A single Ames mutagenicity test costs $500–$2,000 and takes weeks — and that's just one endpoint. Screening 50 compounds across a GeneTox battery means months of waiting and hundreds of thousands of dollars before you know which compounds are worth developing.
Drug discovery teams need a faster, cheaper first filter — to screen large compound libraries computationally, then commit wet-lab resources only to the most promising candidates.
In silico first. Wet lab only when it counts.
Toxometris delivers 50+ ADMET predictions per compound in minutes, at $150 per compound — all endpoints included. Screen your entire library, rank by Risk Score, and send only the top candidates to wet-lab validation.
50+ endpoints. One submission.
Full ADMET coverage across four categories — toxicity, pharmacokinetics, physicochemical properties, and a composite Risk Score.
From SMILES to regulatory report in three steps
Submit your SMILES
Paste one compound or upload a library. No structure drawing required — plain SMILES strings are all you need.
Select endpoints
Choose from 50+ toxicity, ADME, and physicochemical endpoints — or run everything and let the Risk Score rank your compounds.
Get regulatory-grade predictions
Results in minutes. Download a full report with OECD-compliant predictions and QMRF documentation ready for regulatory use.
Built for regulatory-grade use
Not just another ADMET tool. Built to meet the standards your regulatory team actually needs.
All-in-one pipeline
All 50+ ADMET endpoints in a single submission. No juggling multiple tools or stitching results together.
Risk Score ranking
Screen thousands of compounds and rank them by pharmaceutical likelihood — without reviewing every endpoint individually.
LLMs + Graph Neural Networks
Consensus models across SMILES, fingerprints, graphs, and descriptors extract maximum structural information.
OECD & QMRF compliant
Every assay follows OECD 'defined endpoint' principles with QMRF documentation ready for regulatory dossiers.
Cloud-based & always current
Models update continuously as new data emerges — no software to install or versions to manage.
Model fine-tuning available
Need higher accuracy for a specific chemical family? We fine-tune models on your compound class.
Competitively proven accuracy
Our models have been tested against real compounds, by independent scientists, in blind prediction competitions — not just on internal benchmarks.
International AI toxicology blind prediction competition. Competitors had no prior knowledge of the test compounds.
Blind prediction of absorption and fluorescence properties across ~100k EU-OPENSCREEN compounds. Independent validation on a distinct set of endpoints.
Start screening your compounds today.
2 compounds free. No credit card required. OECD-compliant results in minutes.