Skip to content
Vrisko Biosciences

Vrisko Biosciences

AI that turns traditional medicines into drug candidates.

Vrisko-1 is our foundation model for traditional medicine discovery. It predicts structures, designs small molecules and designs proteins.

Public release March 2027.

m/zPREDICTED FITTARGET

Why start with traditional medicine

Evidence is already there. The discovery process needs better tools.

About half

Of the small-molecule drugs approved since 1981 are natural products or are built on one.

More than 100,000

Lives a year saved in Africa alone by artemisinin, a malaria drug that came out of traditional herbal medicine.

30 years

Paclitaxel, from a bark sample in 1962 to FDA approval in 1992.

Sources include Stone et al., 2022; Nobel Assembly, 2015; National Cancer Institute.

View full references
  1. 01Stone S, Newman DJ, Colletti SL, Tan DS. Cheminformatic analysis of natural product-based drugs and chemical probes. Nat Prod Rep. 2022.
  2. 02National Cancer Institute. Taxol: a success story.
  3. 03Bailey CJ. Metformin: historical overview. Diabetologia. 2017;60:1566-1576.
  4. 04Mucuna pruriens in Parkinson disease: a double-blind crossover study. Neurology. 2017;89:432-438.

The Vrisko discovery platform

One loop from traditional medicine to drug candidate.

01Step

Predict and design

Vrisko-1 predicts how the compounds in a medicine fit a disease target, then designs small molecules and proteins from what it finds.

A confidence score with every structure.

02Step

Test

Bioassays on fractions, pure compounds and new designs, chosen by what Vrisko-1 predicts. Activity gets traced to a single molecule, not a mixture.

Results tied to a compound.

03Step

Learn

Every prediction, design and result, failures included, goes back into one dataset for the next round.

Every result kept.

Every result goes back into the next round

Vrisko-1

Vrisko-1 is our foundation model for traditional medicine discovery. It predicts structures, designs small molecules and designs proteins.

01Vrisko-1

Structure prediction

Predicts the 3D structure of a disease target and how a compound from a traditional medicine fits it.

A confidence score with every structure.

02Vrisko-1

Small molecule design

Designs new molecules from compounds that work: more potent, simpler to make, easier for the body to handle.

Nothing is called a candidate until it is tested.

03Vrisko-1

Protein design

Designs proteins that bind a disease target.

Every design is tested for activity before we report it.

Public release: March 2027 · Early access for researchers who want to run it themselves.

Development status

Vrisko-1 is coming in March 2027.

StageStatus
  • Training data

    In progress

    In progress
  • Vrisko-1, first version

    In progress

    In progress
  • Researcher access

    Coming soon

    Coming soon

Last updated 17 November 2026

Early access

Run Vrisko-1 yourself.

We're building Vrisko-1 for researchers to use directly. Leave your details and tell us what you'd try. We'll write when access opens.

Model access for researchers: coming soon

Optional.

Optional.

What would you try first?
What kind of data do you work with?

Optional, but it helps us decide what to build first.

Responsible sourcing

Three promises about how we source.

01

Permission first

We don't collect anything without the permits the law requires and the informed consent of the people who hold the knowledge.

02

Terms in writing

Access, use and benefit-sharing are agreed in writing before any material moves, and the other side keeps a copy.

03

Results come back first

What we learn goes back to the people who made it possible, in a form they can use, before it goes anywhere else.

Partnerships

Partner with us.

Some teams bring us a target and we build a project around it. Others have a traditional medicine they want profiled and tested, a molecule or protein they want designed, data sitting unused, or a need for lab capacity. Tell us what you're trying to find out.