15% of all profit is donated to Heal Palestine and the Palestine Children's Relief Fund (PCRF)
Platform Research Validation Safety Pricing Company Demo Request access
Prospect-1

Drug discovery that starts from the biology.

Prospect-1 is a transparent, safety-gated research pipeline: it selects a target from the genetics, designs candidate molecules, and screens them through ADMET, docking, and physics-based rescoring. Every result is a research hypothesis for a wet lab, never a validated drug.

Current status: research preview. Approved accounts with Prospect-1 and Lab Assistant grants can run the local prospect_design tool for the full in-silico funnel (ranked shortlist, dispositions, AD gating, biosecurity, abstention honesty). There is no hosted production endpoint, no clinical use, and no wet-lab validation claim.
The pipeline

From a target to a testable hypothesis

One path, with two gates that cannot be skipped. Nothing leaves the pipeline without passing a biosecurity screen and a human sign-off.

Screening pipelineresearch use only
Target Design Screen Biosecurity non-negotiable Human check PI sign-off in silico safety gates

A research hypothesis for a wet lab, never a validated drug and never a synthesis instruction. Prospect-1 flags its own compute limits instead of estimating missing values, and abstains rather than guess.

Current evidence

What the prototype can and cannot establish

Implemented research components

Target context, candidate generation, applicability-domain-gated affinity, public-domain ADMET liability signals, hazardous-motif screening, and required human review exist in the research codebase.

Narrow validated scope

Only one commercial-clean target model currently clears the internal exposure gate. Targets without a defensible model abstain instead of receiving a score.

No biological validation

No candidate on this page is claimed to bind in an assay, work in a cell, be safe, or be a drug. Wet-lab confirmation and independent reproduction remain required.

Output contract

A candidate dossier, not a synthesis plan

A reviewable prototype output may contain a candidate representation, the target requested, whether the molecule lies inside a model's applicability domain, an in-silico affinity estimate when allowed, liability probabilities that cleared their own model gates, hazardous-motif flags, and an explicit abstention reason. It does not contain a claim of safety, efficacy, clinical relevance, or a step-by-step synthesis route.

Known limits

  • Small target-specific training sets can make model estimates unstable.
  • In-silico affinity and ADMET signals are not measured biology.
  • License-restricted sources cannot be sold or exposed until legal clearance is documented.
  • Generated chemistry may be novel, infeasible, or outside the domain of every available predictor.
How it is built

Honest by construction

Data integrity first

When the compute or the data cannot support a value, Prospect-1 says so and abstains, rather than filling the gap with an estimate that looks confident and is not.

Biosecurity is not optional

A non-overridable screen runs before anything advances, and Prospect-1 never emits synthesis routes for hazardous chemistry. Safety is a gate in the pipeline, not a setting.

A human closes the loop

Every candidate is a hypothesis handed to a qualified researcher for validation. No result is a recommendation, and nothing reaches a bench without a person deciding it should.

Request research preview access

Prospect-1 runs locally in Lab Assistant for approved accounts. Describe your intended research use, target context, and who will perform any wet-lab review. This is not a request for clinical or treatment advice.