ToxIndex

Smarter data.
Safer chemistry.

Chemical safety assessments backed by curated data, predictive models, and full evidence trails. Get to the science faster.

Everything in one place

Curated datasets, regulatory search, predictive models, and integrations with tools toxicologists already use. Run them through pipelines or just ask a question.

Automated Assessments

REACH dossiers, chemical risk assessments, OEL/HBEL reports, ingredient safety reviews. Provide a substance and get an assessment with an audit trail you can hand to a regulator.

Conversational AI

Ask a question in plain language, get a cited answer. The AI has access to the full dataset and model library, so you can explore data, run predictions, or kick off assessments without leaving the chat.

Knowledge Graph

Hundreds of curated datasets linking chemicals to their hazards, classifications, and study results

Literature

Full-text search across 4M+ regulatory documents from EPA, ECHA, WHO, and 50+ other sources

Predictive Models

ToxTransformer, QSAR Toolbox, and hundreds of other models covering 4,000+ endpoints

Not a chatbot with chemistry docs bolted on

ToxIndex was built from the ground up for computational toxicology.

Reproducible by default

Assessments are designed to be reproducible. Rerun any analysis and get the same result.

Curated, not crawled

We build and maintain the datasets, models, and regulatory databases. Your data plugs in too.

Evidence, not black boxes

Claims cite their source. Predictions cite their model and applicability domain. You can trace anything back.

Built for toxicology

Models trained on tox data and regulatory language, not general-purpose LLMs with a chemistry prompt.

Built for the work toxicologists actually do

Covers REACH, TSCA, SCCS, CLP, K-REACH, and FDA. Export in the format the regulator expects.

Regulatory submissions

REACH dossiers, OEL/HBEL reports, screening-level assessments. Full audit trails, IUCLID export.

Product stewardship

Screen your chemical inventory against global regulatory lists. Flag what needs deeper assessment.

New chemical assessment

No experimental history? Predictive models, read-across, and structural analogs fill the gaps.

Enterprise security

SOC 2 aligned. AES-256 encryption. Workspace-level tenant isolation. Full audit logging. Learn more

Frequently asked questions

What is ToxIndex?

A computational toxicology platform. It combines a curated knowledge graph, regulatory literature search, and hundreds of predictive models to produce chemical safety assessments. You can run full assessments through automated pipelines or explore data through a conversational AI interface.

How does ToxIndex generate a risk assessment?

It runs a pipeline: resolve the substance identity, gather experimental data and regulatory documents, run predictive models to fill data gaps, do weight-of-evidence analysis, and assemble everything into a cited report.

How does the conversational AI work?

It's connected to every tool on the platform. Ask a question in plain language and it figures out the right combination of database searches, model predictions, and literature lookups. You can also trigger full assessment pipelines through conversation.

What regulatory frameworks does ToxIndex support?

REACH, TSCA, SCCS, CLP, BPR, K-REACH, CSCL, and FDA safety assessments. Reports export in framework-specific formats, including IUCLID for ECHA submissions.

Is my data secure on ToxIndex?

Yes. AES-256 encryption at rest, TLS in transit. All data is workspace-isolated with strict tenant boundaries. Hosted on private GCP infrastructure with default-deny networking. SOC 2 Trust Service Criteria aligned.

See it in action

We'll walk you through a live assessment on your substances.