Responsible AI Policy

Nyrava's use of AI is governed by six commitments:

1. Attorney in control. AI proposes. Attorneys decide.
2. Evidence first. Analysis is grounded in your documents, not the model's training data.
3. Transparency. Every engine run and rejection is auditable.
4. Verification. Cited authorities are verified before appearing in drafts.
5. Privacy. Your case content is not used to train general-purpose models.
6. No replacement of judgment. Nyrava does not practice law.

Every factual claim surfaced by Nyrava is anchored to a specific passage in a document you uploaded. The evidence gate suppresses claims that cannot be grounded. Authority citations in motion drafts are checked before they reach the editor.

• AI models can make mistakes, miss context, or produce plausible but incorrect output.
• Automated analysis is a starting point, never a final product.
• Users must independently verify every output against source documents and applicable law before filing or reliance.
• Model behavior can shift when providers update their models. We maintain a router to fall back across providers, but variance is real.

Every output produced by Nyrava — timelines, motions, reports, findings — is designed to be reviewed by a qualified legal professional before use in an actual matter. The product surface makes verification (through inline citations) the default action, not an afterthought.

Nyrava does not use your case content to train general-purpose models. When you configure your own provider API key, requests originating from your workspace are sent directly to that provider under the contractual terms you have with them.

See our Acceptable Use Policy for the full list of prohibited uses.