Paste something you already have — a RAG answer with its retrieved context, or a policy/compliance document — and get back the same kind of scored result the full platform produces. No signup, no install.
The live demo is temporarily paused while the platform is being updated — check back soon.
Paste a question, the answer your AI gave, and the context it was supposed to be grounded in. TrustQE checks whether the answer is actually faithful to that context — not just whether it sounds right.
This covers 3 of TrustQE's 10 RAG checks — the ones scorable from a single pasted example. The other 7 (PII-at-retrieval, injection probes, stale-knowledge detection, answer correctness, and more) actively test your live, connected system, so they need a real endpoint, not a pasted example.
Paste a policy, privacy notice, or any AI-related document. TrustQE checks it against the regulations that apply to your domain and country — EU AI Act, GDPR, DPDPA, and more — and flags what needs a human sign-off.
Paste a question and the answer your AI gave. TrustQE scores it for coherence, fluency, and helpfulness — and if you also paste the source material it was supposed to be based on, it checks whether the answer actually made anything up.
Paste how your AI responded to the same question for two different groups (e.g. two applicants who differ only in a protected characteristic). TrustQE checks both responses for toxicity and leaked identifiers (SSN, email, phone, card numbers, government IDs), and flags whether one group was treated less favourably than the other.
Each check calls your chosen provider directly with your own key — nothing runs on a shared TrustAssay key, and your key is never stored.