Datasets and models, curated for discovery.

Collections you can query.

Genomic variant collections

Population and clinical variant sets with harmonized annotations.

Protein structures and annotations

Experimental and predicted structures with domain and site annotations.

Compound and assay libraries

Screening results linked to compounds, targets and conditions.

Clinical and epidemiological registries

Aggregated registries with documented consent and access terms.

Literature and protocol corpus

Open-access papers, patents and protocols, indexed for citation.

Your own data

Upload or connect. It stays in your project and never trains shared models.

Models trained for research, not for demos.

Three trained models · validation drawn on each card

Variant effect model

Predicts the functional impact of coding variants.

Binding affinity model

Ranks candidate compounds against a target.

Assay QC model

Flags plates and wells that need a second look.

How a dataset enters the workspace.

  1. 01

    Collect

    From a documented source under a clear licence.

  2. 02

    Clean

    Deduplicate, harmonize units and identifiers.

  3. 03

    Version

    Every change gets an immutable version.

  4. 04

    Document

    Provenance, schema and known limits.

  5. 05

    Serve

    Queryable by people and agents, with citations.

A dataset gains its source, version and documents on the way in

Your data stays in your project.

Upload or connect

Files, folders or your own storage.

Project only

Visible only to the roles you choose.

Never trains shared models

Used for your work and nothing else.

Bring your next question.

Tell us your field and what you want to do. We set up the workspace around it.

Request access