Choose your capabilities.
Eleven modules, 5,956 derived properties, 150+ tool API endpoints. Each module has a defined scope, published benchmarks, and clear status. Start with what you need.
Eleven modules, 5,956 derived properties, 150+ tool API endpoints. Each module has a defined scope, published benchmarks, and clear status. Start with what you need.
Mechanism of action and target prediction from a SMILES. MoA predictor, polypharmacology sweep, binding prediction, scaffold library, and the Discovery Wizard in one workflow.
The current ADMET workflow provides permeability, distribution, metabolism and safety predictions. DILI AUROC is 0.952 / 0.989 under the two stated reference policies on the same 96 cases. The current Caco-2 alternative mode gives MAE 0.281 on 182 development-panel cases; the production profile gives 0.914 with test-reference exclusions. Full cohorts and configurations are declared in the endpoint benchmarks.
Program-aware bio-targeting: multi-panel efficacy + human off-target triage + microbiome risk + integrated ADMET. CASF-2016 validated (r = 0.617).
Blind-validated pose prediction — 47.4% top-1 on PoseBusters-428 with no machine-learning training — plus affinity, kinetics and full ADMET / safety integration in one pass.
99.4% accuracy (334/336) on the published 336-case dataset. 7.5 kJ/mol barrier MAE. Typical speedups ~1,000,000× faster than DFT for these tasks.
Operating-window steering for substitution and elimination chemistry with control surfaces, pin/compare, constraint optimization, and audit-ready exports.
Route planning with first-principles barriers. 29 reaction types at 3.1% MAE. 200 specific reactions at <1% MAE. 82 reagents, 15 disconnection patterns.
Predict spectra before synthesis. UV-Vis 5.4% error, IR <1%, NMR 0.3–0.5 ppm MAE.
Marcus theory + first-principles tunneling corrections. 26 / 26 tests passing. Through-bond decay constant matches literature ranges. 2–3× tunneling enhancement.
Solution-phase accuracy. 0.71 logS MAE on 1,128 ESOL compounds. 45 solvents implemented, 6 benchmarked.
MD simulation with trajectory viewer, plus Gibbs / Helmholtz free energy and temperature-dependent thermodynamics with AI interpretation.
Ionization-state characterization and long-range interaction coefficients. pKa with automatic ionization-site identification; C6 dispersion coefficients from polarizability.
A suite of advanced chemistry methods for demanding problems: excited-state spectroscopy, QM/MM multi-scale calculations, photochemistry, relativistic corrections, and complete-basis-set extrapolation.
Carry crystalline-silicon material state through optics, transport, recombination, junction behavior, and contacts to the terminal J-V quantities device engineers use.
Composition-only band-gap predictor with the full electronic-property bundle: alloy decomposer, dopant classifier, mobility, conductivity, band-edge alignment for heterojunctions. ~1 ms per query, no machine-learning training.
Real-time drift mobility across doping, temperature, and alloy composition. Design targets with pass/fail overlays, inverse spec search across a 34-material library, Pareto frontiers, (N, T) operating-box heatmap, and CSV export.
Catalyst-native scoring across eight industrial families with full cycle simulation, microkinetics, electrocatalysis (potential + pH), and constraint-driven inverse discovery. DFT-competitive barriers at microsecond speed.
Battery-native decision layer for cathodes and build-ready screening. Combines transport, degradation, electrolyte & coating fit, uncertainty, and prototype hand-off in one workflow.
Work function, contact readiness, and interface engineering. Surface-aware scoring for device stacks — metal/semiconductor junctions, Ohmic vs Schottky, heterostructure alignment.
Generative materials discovery: stream candidate compositions that satisfy a multi-objective spec in real time. Watch the pool evolve live as constraints tighten.
Iterative lead-compound optimisation: evolve thousands, GPU-validate, promote the lead, repeat. Survival-weighted operators, repair-first moves, and generational diversity guards.
Define what you need. Get candidates that match. See why others were rejected. Works across life-science, chemistry, and materials workflows.
Projects, versioned runs, pipelines, notebooks, and comparison views. Answer “why did we pick this?” six months later.
Run any property across hundreds or thousands of molecules in one submission. Bond lengths, ionizations, pKa, druglikeness, NMR, materials, formation enthalpies, interactions — all batch-ready.
Central storage for your molecules, materials, and candidate sets. Tag, version, and pull into any workflow — used by Batch Processing, Inverse Search, and every specialty module.
Most pilots start with 1-2 modules. Tell us about your workflow and we'll recommend a starting point.
Request Pilot Access