Pricing

Priced per program, not per seat.

Early access pricing for teams running active protein engineering programs. The Explorer tier is free and exists because you should know whether the model works for your target before paying anything.

Pricing tiers

Explorer

Free

Free retrospective validation run

One held-out retrospective validation on a sequence-activity dataset you provide. Determines whether the model discriminates meaningfully in your fitness landscape before you pay anything.

  • 1 retrospective validation run (up to 200 sequences)
  • Rank correlation report delivered within 5 business days
  • Email Q&A with Elena or Marcus
  • No commitment required
Request validation

Lab

$1,200/mo

per active program, per month

One active protein engineering program. Unlimited generative search runs. Ranked shortlists with residue-level annotations. Results in CSV and FASTA format.

  • 1 active program
  • Unlimited search runs within program
  • Ranked shortlist FASTA and CSV delivery
  • Per-residue mutation impact annotations
  • Retrospective validation included
  • Email support, response within 1 business day
Get access

Multi-Program

Custom

Contact us

3 or more simultaneous active programs. For organizations running parallel therapeutic and enzyme programs, or research groups with multiple distinct protein engineering objectives active at the same time.

  • 3 or more active programs
  • Dedicated onboarding call
  • Priority search queue
  • Custom output formats on request
  • Quarterly model update briefing
  • Slack or Teams channel support
Talk to us

Questions before you commit

A program is a single protein engineering objective: one target protein with one defined fitness goal. Running CDR optimization on Antibody A and thermostability work on Enzyme B are two separate programs. Running CDR optimization across multiple antibody candidates with the same scaffold and objective is one program.
Yes. The Explorer tier is specifically designed for this. You provide a held-out set of sequence-activity data, we run the model, and you receive a rank correlation report within 5 business days. No payment required. If the model does not discriminate meaningfully for your target, we tell you.
No. Data you provide for retrospective validation or program runs is used only for your program. It is not included in base model retraining. The base model trains only on public sequence databases. We maintain strict data separation between customers.
Coverage depends on how well-represented your protein family is in public databases. Kinases, immunoglobulins, oxidoreductases, proteases, and transporters have strong coverage. Orphan enzymes or recently characterized families may have thin coverage. The retrospective validation is the honest way to check before committing.
Ranked candidates are delivered as a FASTA file and a per-residue score CSV. The CSV includes predicted fitness score, per-residue mutation impact annotations, and a model confidence note for your protein family. Multi-Program customers can request custom output formats.
The Explorer tier is the free entry point. It is not a time-limited trial with features artificially removed. It is a meaningful retrospective validation on your actual sequence-activity data that tells you whether the model discriminates for your target. We think that is more useful than a 14-day sandbox. If the retrospective does not show useful rank correlation, we will say so and you will have lost nothing except the time it took to prepare a held-out dataset.

Not sure where to start?

Tell us your target protein and fitness objective. We will tell you honestly whether Explorer makes sense or whether your coverage question can be answered in an email first.

Talk to us first