Claim assessment
Check a specific factual claim, headline, post, or assertion. Evidrai returns a verdict, confidence, caveats, and source trail.
Product
Evidence checks for claims, speeches, articles, and public narratives.
Evidrai helps early-access users separate evidence from repetition. It is built for people who need to inspect why a claim looks supported, contradicted, weak, or still unresolved.
Speech and video review
Extract checkable claims from a transcript or supported video URL, then choose which claims deserve verification.
Saved evidence reports
Assessments are saved to your account so you can revisit evidence, feedback, verdicts, and source scoring later.
Enterprise Gateway Mode
A verification control point for AI-generated outputs in consequential workflows. Each output gets a pass, review, or block decision — evidence-backed, policy-driven, and auditable.
How it works
Show the work, not just a verdict.
Each assessment is designed to expose the reasoning path: what was checked, which sources mattered, how strong the evidence was, and what uncertainty remains.
Evidence beats repetitionConfidence is not certaintyPrimary sources matterCaveats stay visible
Evidrai Enterprise
A verification layer for AI workflows that carry real consequences.
Enterprise Gateway Mode puts Evidrai between your AI system and the actions it would trigger. Every decision is policy-driven, evidence-backed, and written to an audit log. Pass, review, or block — inline, in one API call.
Pass — evidence supports the outputReview — route to human verificationBlock — evidence contradicts or is absentFull audit record on every decision
Available now
- Standard claim checks with evidence review
- Article/source URL context
- Pasted transcript speech audits
- Best-effort YouTube transcript extraction
- Saved reports and feedback capture
- Evidence scorecards and source grouping
- Polished report export and share links
- Enterprise Gateway Mode (early access)
Coming next
- Researcher evidence ledger
- Durable source snapshots
- Standard API access with keys and usage limits