Rasayan Labs Launches AI-Native Drug Discovery Platform at ICT Mumbai

San Jose- and Nagpur-headquartered startup Rasayan Labs has launched Rasayan.AI, an AI-native computational chemistry and drug-discovery platform, at the Institute of Chemical Technology (ICT) Mumbai. The system is designed to bring several tasks that researchers often handle through separate software tools into one workflow, including synthesis planning, molecular-property prediction and chemical analysis.

The platform combines AI-powered retrosynthesis, ADMET prediction across more than 1,400 properties, impurity prediction and RasayanDraw, the company’s molecular editor. Rasayan says its underlying datasets span more than 13 million chemical reactions, 126 million building blocks and 110 million patents. These scale figures are company-reported and do not by themselves establish drug-development performance.

Rasayan has also tested part of its modelling stack through the OpenADMET PXR Induction Blind Challenge. The company reports that its final model ranked ninth on a blinded 260- test set, placing it in the benchmark’s top statistical tier. PXR prediction matters during early drug discovery because activation of the receptor can affect pathways involved in drug metabolism and potential drug-drug interactions.

The launch builds on a formal research collaboration between Rasayan Labs and ICT Mumbai. According to the organisations, work on dual COX-2/mPGES-1 inhibitors produced 23 proposed candidates in roughly three weeks, with 19 moving into synthesis. That result represents a design and early research milestone, not clinical validation or an approved medicine. Academic researchers and students are being offered free access to the platform, while enterprise deployments can be configured for cloud, hybrid, on-premise or air-gapped environments.

The broader opportunity is reducing the friction between molecular design and laboratory experimentation. AI can narrow chemical search spaces and prioritise compounds, but promising computational predictions still require synthesis, experimental testing and eventually extensive preclinical and clinical validation. Rasayan’s next challenge will therefore be demonstrating that a unified AI workflow consistently improves research productivity when used in real pharmaceutical programmes.