How PyLabs Works
An AI prediction layer connecting molecular candidates with their physical outcomes before any synthesis or TD-DFT validation.
Molecular AI prediction layer
By framing molecules as graphs and using advanced Graph Isomorphism Networks (GIN) trained on high-fidelity, solvent-aware experimental and synthetic datasets, PyLabs maps SMILES strings directly to photophysical consequences.
The model evaluates absorption, emission, bandwidths, extinction, quantum yield, and lifetime instantly.
Built for the R&D workflow
We do not replace your existing validation infrastructure - we dramatically improve its hit rate. Whether you are working with a virtual library of a hundred compounds or running a ten million candidate screen, PyLabs filters down to the top percentile.
- Batch Screening: Feed extensive multi-scaffold CSVs and receive predictions matrixed across different solvents.
- API Workflow: Connect your proprietary generation tools directly to our prediction endpoints.
- Relationship to TD-DFT: Use PyLabs to eliminate non-starters, then spend your quantum-chemistry computational budget on the most promising candidates.
Evaluate our baseline.
Test the platform against previously characterised molecules from your internal records.
Request Access