The pilots

Send us the lot
that's fighting you.

The isolate, hydrolysate or candidate set that keeps reading bitter, beany or off. We read each compound from structure and hand back the call — and you check it cold against your own data. Three blind pilots; pick the read you need. It's the path to a partnership.

Send us your lot ← Back to the platform

Pick the read you need

Three pilots.

The format you send is the read you've subscribed to — three separate pilots, each correlated blindly against your own panel and lab.

Pilot 1

Taste

You send name + SMILES

The taste call, whether it's an off-note, the mechanism that fired, the ranked masker to try first, and blend palatability.

Pilot 2

Product-readiness

+ storage, solubility

Everything in Taste, plus aroma (beany / green / rancid) and shelf-life — oxidative and light (photostability → packaging).

Pilot 3

Full dossier

+ dose, genotype

Everything above, plus the full safety stack (reactive warheads, dose-aware DILI), bioavailability, and genetic perception (supertaster vs non-taster).

Every pilot is blind — nothing is fit to your compounds, so a correct call is a genuine prediction, not a lookup. It's a pre-bench prioritizer, not a replacement for your panel or assays.

Why you can trust it: the engine was correlated against a University of Manitoba lab — bitterness predicted from structure vs an electronic tongue, R² 0.95 across 20 samples. The pilot proves it again, on your compounds.

See the evidence →

Where it leads

A path to partnership.

Bring us the lot that's costing you.

You get the answer; we get the proof that the method is real on your own data. The pilot proves it; the partnership is where we take it further together. One door opens the rest.