Simple tool for predicting solubility of small molecules.
This site provides free access to the fastsolv solubility prediction model described in DOI |10.1038/s41467-025-62717-7|.
Contact jwburns|@|mit.edu or lattia|@mit|.edu for any issues.
To make predictions, enter solute SMILES in the field on the left and solvent SMILES in the field on the right, then finally enter your desired temperature range. Click Run Prediction to generate results, which are then made available in a Comma Separated Values (csv) format. To install fastsolv locally for faster inference and more customization, see github.com/jacksonburns/fastsolv.
Please cite the companion paper 10.1038/s41467-025-62717-7 if you use these predictions in published scientific work. Citation: Attia, L., Burns, J. W., Doyle, P. S., & Green, W. H. (2025). Data-driven organic solubility prediction at the limit of aleatoric uncertainty. Nature Communications, 16(1), 7497.