Generative
Chemistry

ADMET &
Off-target

Retrosynthesis

Model
Training

MDFlow

Alchemistry

A Diverse Toolbox for Drug Discovery

Hit Identification

De novo generative chemistry,
Virtual Screening

Hit-to-Lead

R-group exploration, Scaffold hopping

ADMET & Selectivity optimization, Relative Binding Free Energy estimation

Lead Optimization

Launch a Ligand-Based or Structure-Based Drug Discovery workflow.

Generative Chemistry

Generative models ensemble will Create optimal molecules based on your criteria.

Prioritize results using Alchemistry, ADMET profiling, and Golden Cubes.

Customize the platform with your external tools such as QSAR models, MD simulators, and in-house databases.

Predict reliable synthetic routes for uploaded or generated molecular structures.

Retrosynthesis

Trained on the unique library of Expert-annotated Reaction templates with a focus on the medicinal chemists’ reaction toolbox.

Collection of 300K Commercially available Building blocks
indexed within searchable CAS registry numbers.
Extensive support of chemo-, regio- and stereo-selectivity.

Our AI models are trained to perform well on novel chemotypes, allowing you to confidently use them in generation and screening experiments.

Model Training

Train a state-of-the-art model on your data, including in vitro activity or simulation data.

MDFlow

End-to-end simulation workflows for biomolecules and complexes, from system build and sampling to reproducible reporting and metrics.

Physics-based core — an algorithm built on fundamental principles of physics for unprecedented accuracy and speed

Alchemistry

Compute the relative binding free energy between your protein and ligands to prioritize your lead compounds. Augment your analysis with experimental data to determine the absolute binding free energy.

ADMET & Off-Target

Predict and optimize the ADMET and selectivity profile of your molecules. Use our predictive models as a standalone profiling tools, or within the Generative Experiments to guide the design toward future drug candidates.

Golden Cubes

Golden Cubes is a proprietary engine that achieves high accuracy thanks to multi-dimensional Self-Organizing Maps.

Predict off-target kinome selectivity.
Golden Cubes works with 2D and 3D structures, using models trained on carefully curated activity datasets.