Examine the material your AI relies on.

Check documents and structured data

Select the checks relevant to your source collection. Findings preserve their supporting material and distinguish contradicted, unsupported, and unresolved claims.

• Compare claims within and across related documents
• Check a finished document against supporting sources
• Detect inconsistent structured values, missing cells, and missing record keys
• Identify duplicates, competing versions, and unreadable material
• Review examined scope, exclusions, and cases requiring human judgment

Evidence you can retain, share, and inspect independently

Successful, seal-eligible paid audits produce readable reports and machine-readable findings, with recorded methods, scope, and limitations. A signed manifest and portable checker let recipients detect altered, missing, or added files without a VeriTrooper account or installation. Human signoff is included when recorded. Evaluation provides a limited proof set; companion documents depend on the selected checks.

• Bound Data Audit Report and Executive Summary
• Audit Scope, Data Card, and Methodology records
• Detailed findings, source support, and review referrals
• Suggested corrections and remediation priorities where applicable
• Framework Crosswalk and applicable Annex IV Data Record
• Evidence JSON, findings JSONL, data-card YAML, SARIF, and custom evaluation records
• Canonical results, configuration, signed manifest, and portable checker

Understand what coverage means

Structured coverage measures completeness within the records and fields examined. Prose consistency checks have a recorded scope and comparison budget. Neither an empty findings list nor full examination of one population proves that the whole collection is correct or complete.

Run where the data lives

Use the bundled local model and deterministic checks for local operation. Retain findings on your own storage and verify package integrity offline. Optional remote model roles send the required material directly to the providers you select.

One lifecycle, separate decisions

SitRep examines source material. Scout tests AI answers. Watchtower reviews deployed answers over time. The shared Governance Records workflow supports accountable human review of their evidence.