Predict chemical reactions, retrosynthesis pathways, and experimental procedures. Train teachable AI models that can build intelligence in specific chemistry domains. Scale your analysis and model training while securing your data using features of the Discovery Platform. Execute chemical synthesis procedures in an autonomous laboratory.
Our tool is based on molecular transformer models that understand the natural language of chemistry, trained on 2.5 million chemical reactions. These models are flexible, adaptive to new data and non-rule-based.
It is part of the Discovery Platform, using specialized infrastructure to train multiple complex AI models simultaneously and with greater scalability. Users never have to limit training times or wait in queues, resulting in more accurate models, created faster, with better uptime.
Run predictions and train models at scientific scale without worrying about provisioning and monitoring backend resources
Integrate into your own platform-based workflows with minimal setup. Use the API to access predictions in the same view as other platform features.
AI models generate results without human bias, opening up nore discovery possibilities. Models can be made more intelligent with additional training data you provide.
A visual drawing editor (based on Ketcher) allows you to design your own chemical structures, and retrain AI models without needing to code
Preconfigured libraries of molecules allow you to add reactants and reagents to your Ketcher board with just a few clicks.
Create projects and share them with friends or colleagues. Collaborate on complex multi-step reaction syntheses or on novel chemical reaction designs.