Predictive safety intelligence for chemicals and advanced materials
QSAR Lab's two prediction platforms, under one roof. Thirty-five validated models — documented to OECD principles, with the applicability domain checked on every prediction — so you can screen more substances against more endpoints without building a modelling team.
One system. Two domains.
A molecule and a nanoform are not the same kind of object, and they do not take the same kind of model. QLsystems keeps both under one brand, one access model and one documentation standard — while letting each platform be genuinely native to its domain.
Different inputs, by design
QLchemicals starts from a molecular structure. QLmaterials starts from measured nanoform characteristics — particle size, shape, crystal structure — because a nanoform has no single molecular graph to hand a model.
The same documentation standard
Every model in both platforms carries a QMRF built around the OECD validation principles: defined endpoint, unambiguous algorithm, defined applicability domain, and separately reported fit, robustness and predictivity.
The same access model
Token-based. No subscription, no software to install, no chemoinformatics expertise assumed. One prediction costs one token.
Honest about limits
Most of these are local models. The catalogue reports small validation sets and synthetic oversampling where they exist, because a screening tool that oversells itself is worse than no tool.
Endpoints in the platform
Four model families
From structure to documented prediction
The six TiO₂ endpoints
Scope
From material definition to a QPRF report
Every model, in the open
All thirty-five models with their real statistics — including the awkward ones. Filter by platform, family, model type or computational cost; open any model for its full specification, applicability domain and source.
One prediction
1 token
per model prediction
from 9.99
per token · academic discounts on request
Ask about academic pricing