Shortlisting materials before making them
Could we recognize a good material before making it?
For a material with a set job, such as a battery or a catalyst, we think the candidates are endless while the few you can actually make and measure are limited. We want to shortlist candidates from molecular structures and property data first, so only those worth testing reach the lab. Sharing our criteria and evidence would let other labs use them too.
Molecular structures and property data
We aim to connect and clean the data, then have agents compute properties and verify each other.
A ranked list of candidate materials, with evidence
| Field | Chemistry |
|---|---|
| Status | Planned |
| Input | Molecular structures · property data |
| Output | Ranked candidate materials |
| Delivery | Joint research |