Technology
onepot C1 — our first generative model
C1 is a synthesis data-aware generative model. Given a reference molecule, substructures to preserve, or property targets, it proposes makable analogs with forward routes from commercially available building blocks. And we can synthesize these analogs.
Conditional sampling
Anchor generation on a molecular backbone, substructures to preserve, property windows, or any combination of those modes in a single run.
Chemically feasible by design
C1 does not just generate plausible structures. It prioritizes molecules with routes through chemistry that is feasible in the lab.
Customizable to your program
C1 can be adapted to project-specific scoring models, property filters, and design rules, so it can propose molecules optimized for the signals your team already trusts.
Behind Designer and the API
C1 powers the Designer UI and the onepot API for programmatic use.
C1 starts from a reference backbone, branches into analogs, and selects feasible candidates optimized for project-specific scoring models.
Backbone
reference
Analog A
substructure
Analog B
score +0.34
Variant 19
score +0.63
Variant 27
score +0.71
Candidate 11
score +0.91