Five capabilities. One governance runtime.
Most AI governance tools are dashboards. Constrix is infrastructure — inline, non-bypassable, and active before your AI executes. Here is what it does.
Every AI decision evaluated before it executes
Constrix sits inline in your AI pipeline. Before any agent acts, it passes through eight evaluation stages — schema validation, context loading, policy compilation, OPA evaluation, and cryptographic sealing. One of four decisions is returned in under 5ms: allow, restrict, deny, or kill. No stage can be skipped. No decision is silent.
Control what your AI agents are allowed to do
AI agents connect to tools and data sources via Model Context Protocol (MCP). Without governance, agents can call any tool with any parameters at any time. Constrix evaluates every MCP tool call before it executes — approving, restricting, or blocking access based on your governance rules — and seals the outcome.
Know exactly which AI agent did what — and when
Every AI agent deployed through Constrix receives a cryptographic identity — a short-lived Ed25519-signed AIT. Actions are attributed to a specific agent, session, delegation depth, and timestamp. Ghost agents — dormant but credentialed — are automatically detected and revoked. The complete lifecycle is sealed in CAPL.
Turn your regulation into working governance rules — automatically
ComplyAI maps regulatory articles from EU AI Act, ISO 42001, NIST AI RMF, and SOC 2 to Rego governance policies that Constrix enforces inline. Regulatory alignment also available for GDPR, HIPAA, and Saudi SDAIA. When an auditor asks for proof, ComplyAI generates a signed evidence package — not a report, not a log, but cryptographically sealed records verifiable without trusting Constrix infrastructure.