Maetra gives engineering, security, and compliance teams one control plane to discover agents, classify risk, route approvals, block unsafe actions, and prove every decision with an immutable audit trail.
Each module works alone. Together, they compound. Discover finds the agent. Comply classifies the risk. Govern applies the right approval policy. Secure checks the action in real time. Audit proves what happened.
Maetra scans GitHub repos and connected sources to identify official, shadow, and unreviewed AI agents. Every pull request becomes a governance checkpoint.
Classify AI systems across EU AI Act, GDPR, NIST AI RMF, ISO 42001, Colorado AI Act, and SOC 2. Generate documentation from the same underlying agent record.
Agents call Maetra before acting. Low-risk actions are logged. High-impact actions trigger role-based approvals, quorum policies, timeouts, and escalation paths.
Scan prompts and tool calls in real time for prompt injection, data exfiltration, privilege escalation, and policy violations before they reach production tools.
Maetra records every discovery, classification, approval, rejection, policy change, and security incident in a hash-chained evidence log.
Developers see governance impact before code merges. Maetra comments on AI-related changes and can block merges until review is complete.
Agents call Maetra at decision points. Maetra returns approved, pending, or blocked with a signed decision token.
const decision = await maetra.checkpoint({ agent_id: "finance_agent_prod", action: "transfer_funds", payload: { amount: 84000, currency: "USD" }, timeout_seconds: 14400 }) if (decision.status === "approved") execute() if (decision.status === "pending") awaitWebhook() if (decision.status === "blocked") stop()
Connect GitHub for PR governance, Slack for approvals, Email for fallback, Webhooks for custom workflows, and SIEM for enterprise audit streams.
Start free with a single module. Upgrade when you need enforceable controls, more agents, and regulator-ready evidence.
For small teams exploring AI governance.
For growing teams with active AI agent deployments.
For regulated organisations with compliance requirements.
Start with one module. Expand into a full AI governance control plane.
Suggested autonomy: L3 → L4
EU AI Act: high-risk employment workflow
Action required: update approval policy before merge
Transparency notice and Colorado consumer disclosure may be triggered.