verifiable-ai-governance

Verifiable AI Governance is described as a vendor-neutral reference platform for turning AI governance requirements into deterministic controls, independent approvals, signed authorization, runtime enforcement, governed response and verifiable evidence. The project is designed for organizations that need to answer not only what a policy requires, but also what was authorized, what happened at runtime, which control was applied and which evidence proves it. The current release-candidate code connects governance decisions to runtime evidence through an explicit chain: Policy → Approval → Signed Authorization → Runtime Enforcement → Violation / Runtime Assurance → Governed Response → Evidence. The API is authoritative for state transitions, authorization, segregation of duties, versioning and audit; frontend validation is never treated as a security boundary. Maturity is a functional, production-oriented reference implementation, with selected enterprise integrations still requiring validation in real organizational environments. A public read-only demonstration is available and contains synthetic data; the repository deliberately separates reference proof from deployment claims.