Owner-approved access

Plans

You pay for answers, not compute. Every engagement starts with a $1,299 pilot: you pick the proteins, withhold the measured numbers, and score our sealed predictions against them yourself.

Start with a pilot

A blind test on things you have already measured. You set the success bar with us first, we predict, then you reveal your numbers and score us.

See the method on public benchmarks first →

The pilot
$1,299 CAD, one time

A deliberate first-look price: we price the pilot to get the method onto your bench, and earn our rate only if it becomes a project.

What a pilot covers
  • Up to 50 small molecules, a flavour compound, a sweetener, an off-note, each comes back with its taste and aroma read.
  • Or up to 10 proteins, each comes back as a full instrument dossier, so one protein is a much bigger job than one molecule.
  • Send things you have already measured (a mix is fine, we agree the scope first), so you can score us against your own numbers.
What you get back, in three business days
  • For a protein: a full instrument dossier, melting point, gelation, solubility, allergen family and more, so you know how it will behave before the bench.
  • For a small molecule: the taste or aroma call, how strong it is, and any off-note.
  • An identity mass-check, included free.
  • Where a call is not reliable, it says so instead of guessing.
How you score us, agreed up front
  • You set the bar first, we agree what counts as a hit before you send, matched to your data.
  • We send predictions first, timestamped; you reveal your measured numbers afterward and score us yourself.
  • If a call is genuinely uncertain we abstain rather than pad the score (capped at 20%).
  • The bar: at least 70% correct against what we agreed.

Every prediction is deterministic, re-run it and the number is identical. For protein melting points we score against a disclosed accuracy band (about 3°C within a known fold family), not a single temperature, since the same fold naturally varies by species; the full method is on the benchmarks page.

Your data
  • Under NDA, used only for your pilot, deleted on request, never used to train anything.
  • Clear the bar and go ahead, and your $1,299 comes off the first project milestone.
Request a pilot

After the pilot

If the pilot earns it, you choose how to go on, priced to what you need.

Scoped project

A project you can hold us to. The pilot becomes a milestone project, each milestone accepted against criteria set in writing before it starts. Your pilot fee comes off the first milestone, and the scope is your candidates.

Platform access

Your team runs it directly, for an agreed volume of inputs. Named users submit structures and pull sealed dossiers up to that volume, and the in-domain list grows as we validate new families. You keep every result, no royalty. Priced to the volume you need.

A screen, not an assay. RaRaMa front-runs the characterization you would otherwise pay for one assay at a time, so you spend bench budget only on the candidates that earn it. We tell you which are worth your bench and your panel, the lab confirms the winners.