QSAR Toolbox download and use for REACH hazard assessment
The OECD QSAR Toolbox is free software for chemical hazard assessment. Find the official download, first workflow and conditions for REACH read-across.
The OECD QSAR Toolbox is free chemical hazard-assessment software for finding analogues, grouping chemicals and filling data gaps. It supports read-across, trend analysis and QSAR models, alongside experimental data retrieval and chemical profiling. The official Toolbox site describes these functions as support for transparent, reproducible assessments.
For a REACH dossier, the software output is only part of the evidence. ECHA requires a scientific justification for the prediction or read-across used to meet an information requirement. A generated report does not establish acceptance. See ECHA's guidance on QSAR models and grouping and read-across.
QSAR Toolbox download and training
Use the official QSAR Toolbox download page, rather than a third-party installer mirror. Its installation instructions tell users to decompress the archive first. Start installation from the extracted folder, not from inside the compressed file.
The Toolbox is free of charge under an End User License Agreement accepted during installation. That does not mean unrestricted licensing. It is a decision-support system for people with sufficient toxicology or ecotoxicology knowledge, so expert judgement remains part of the work. The official overview explains its intended use and licence basis.
For QSAR Toolbox training, start with the official support page. It provides getting-started guidance, video tutorials and ECHA webinar links. It also covers category-consistency assessment, prediction reporting and exports.
Check the tutorial's version before following its instructions. Some official tutorials cover earlier releases, so their instructions may not match your installation. Use guidance that matches your installed release where available.
A first QSAR Toolbox workflow
The official tutorials follow a sequence from target selection to reporting. Use this as an orientation, rather than a fixed procedure for every endpoint:
- Enter the target chemical and define the endpoint. Keep the assessment focused on the property you need to address.
- Profile the chemical. Review the information used to describe its structure and relevant mechanisms.
- Collect available data. The Toolbox can retrieve experimental data as part of the assessment.
- Search for analogues and define a category. Consider structural and mechanistic similarity when selecting source substances.
- Select a gap-filling method and refine the category. The tutorials cover read-across, trend analysis and QSAR approaches.
- Report and export the assessment. Follow the relevant reporting guidance for the method used.
The category needs a toxicological basis. A set of similar-looking structures does not, on its own, explain why source-substance data apply to the target. The Toolbox overview describes structurally and mechanistically defined analogues and categories as central to its approach.
Do not treat the application as one universal toxicity model. The Toolbox site lists profiling, metabolism simulation and external QSAR models among its functions. The method you select still needs to fit the assessment question.
QSAR predictions and read-across under REACH
QSAR and read-across describe different ways to fill a data gap. ECHA defines QSARs as mathematical models that predict properties from chemical structure. Read-across uses relevant information from analogous source substances to predict a target substance's properties. The Toolbox supports both, but the terms are not interchangeable.
For QSAR use, ECHA sets out conditions covering the model, target substance, result and documentation:
- The model must be scientifically valid.
- The target substance must fall within its applicability domain.
- The result must be reliable and sufficient for the relevant information requirement.
- The documentation must allow independent evaluation.
ECHA also recommends thorough documentation in standardised IUCLID formats. These conditions concern the evidence behind an individual prediction, not simply whether a particular software package produced it. Follow ECHA's QSAR guidance when assessing a result for dossier use.
For read-across, ECHA calls for a hypothesis-driven justification specific to the property being assessed. Address substance identity, composition and impurities. Explain both structural similarities and differences, examine contradictory experimental evidence, and provide a category data matrix. ECHA recommends its Read-Across Assessment Framework to check the argument's strength. Its read-across guidance identifies REACH Annex XI, section 1.5 as the conditions for this adaptation.
A Toolbox assessment can support a REACH adaptation when the relevant conditions are met. ECHA requires read-across justification for each applicable standard information requirement. It assesses whether the registrant's justification meets the legal requirements, so a software report alone is not enough.
The Toolbox support page lists the QSAR Assessment Framework, QSAR Model Reporting Format and QSAR Prediction Reporting Format. Use these resources for assessment and documentation, without treating a completed format as proof that a result will be accepted.
Where Japan's HESS fits
Japan's HESS, the Hazard Evaluation Support System Integrated Platform, has a more specific stated focus: repeated-dose toxicity evaluation using a category approach. NITE's HESS page, updated on 12 June 2026, describes its toxicity and metabolism knowledge. This includes toxicity mechanisms, rat metabolism maps and ADME information for rats and humans.
NITE describes HESS as compatible with the OECD QSAR Toolbox. Choose HESS resources for their repeated-dose assessment scope. Check NITE's instructions before planning data transfer between the systems.
Access is separate. NITE offers HESS and HESS DB through free user registration. Download instructions follow in the registration-verification message. Its page describes regulatory use under Japan's Chemical Substances Control Law as expected. Do not treat this as blanket acceptance of every assessment.
RegAffairs AI's answer on the Toolbox, REACH and HESS provides a regulatory research starting point for these questions.