AI Decision Governance — Pre-propagation, not post-mortem.
Portotify is a deterministic governance boundary for consequential decisions. Every decision — whether produced by an AI model, a rule engine, or a human — is evaluated before it propagates.
The protocol decides: ALLOW, BLOCK, or REVIEW_REQUIRED.
- Fail-closed: unknown states produce blocks, not silent passes
- Deterministic: every verdict is reproducible from rules and inputs
- Vendor independent: same governance runs over OpenAI, Anthropic, Google, or in-house models
- Immutable audit trail: append-only, versioned, tamper-evident (SHA-256)
- EU AI Act ready: framework mapping for Annex III articles, full enforcement 2 August 2026
→ PORTOTIFY_DEMO_PUBLIC — recorded ALLOW, BLOCK, and REVIEW outcomes with production latency benchmarks