AI Agent Guardrails API Documentation

Runtime guardrails, AI passport verification, and policy enforcement for production AI agents. Secure OpenClaw, LangChain, and CrewAI with prebuilt guardrail policies.

🚀 Quick Links

🌐 Website 💻 GitHub 📚 Guides 📄 OpenAPI Spec

🔐 Authentication Methods

✓ No Authentication

Public verification endpoints - no credentials required.


              GET /api/verify/*

🔑 API Keys

Server-to-server access with API key authentication.


              X-API-Key: apk_...

🛡️ JWT Tokens

User session tokens for interactive applications.


              Bearer eyJ0eXAi...

💡 Pro Tip: Click the Authorize button at the top to configure authentication for "Try It Out" requests. Your credentials will be persisted across page refreshes.

🛡️ AI Agent Guardrails & Policy Enforcement

APort provides production-ready AI agent guardrails with 20+ prebuilt policy packs for finance, data, code, system, web, and communication operations. Enforce guardrails before the API call with sub-100ms verification.

Finance Guardrails

Payment limits, refund controls

System Guardrails

Command execution, tool safety

Web Guardrails

Fetch controls, browser safety

⚡ Getting Started with AI Agent Guardrails

Create AI passport - Establish cryptographic agent identity
Attach guardrail policies - Use prebuilt policies from /policy-packs
Verify before execution - Call /api/verify/policy/{pack_id}
Get cryptographic proof - Receive court-admissible attestations

✨ New to AI Agent Guardrails? Check out our resources:

Quickstart Guide Builder (Visual) Policy Packs OpenClaw Guardrails