Veyra Atlas API opens to external partners
The Software & AI Services division has made the Veyra Atlas materials-property prediction model available to external clients via a documented REST API. Atlas, which was trained on 14 years of characterization data from the Institute's core facilities, predicts thermal conductivity, elastic moduli, and optical bandgap for amorphous and crystalline compounds with a reported mean absolute error of 4.7% against experimental reference sets.
External access is offered under a tiered subscription structure. Academic partners may apply for a research-use licence at no cost, subject to publication-sharing terms. Commercial clients access the API through the Software & AI Services rate schedule. Trial access requests can be submitted to atlas@veyra.example. Lead developer Dr. Naila Ravelo will present the API capabilities at the June all-hands seminar and at the upcoming QPS/CDS joint workshop in September.