Cookbooks

Cookbooks

Practical, step-by-step guides for integrating Checkpoint into your applications

Welcome to Checkpoint Cookbooks

Cookbooks are hands-on, scenario-based guides that walk you through specific integrations from start to finish. Unlike general documentation, cookbooks focus on doing — with real code, real configurations, and real outcomes.

New to Checkpoint? Start with Core Concepts for the mental model, then return here for implementation guides.

Deploy your first AI agent with its own cryptographic identity in under 10 minutes. Connect it to Telegram and start chatting.

Choose Your Path

Checkpoint offers three complementary capabilities. Not sure which fits your integration? See Choose Your Integration for the canonical decision guide. Otherwise, pick the capability that matches your goal:

Detect — Know What's Visiting Your Site

Passive AI agent detection for analytics and visibility. No blocking, just intelligence.

Enforce — Block Unwanted Agents

Active detection plus policy enforcement. Block, redirect, or challenge detected agents.

Govern — Authorize AI Agents

Deploy KYA-OS servers to grant controlled access to verified AI agents.

Quick Decision Matrix

Not sure where to start? Use this matrix:

I want to...UseCookbook
See what AI agents visit my marketing siteDetectPixel
Track agent activity in my web appDetectBeacon
Block ChatGPT from scraping my contentEnforceGateway
Protect my API from unauthorized agentsEnforceMiddleware
Let verified agents use my MCP serverGovernDashboard Deploy
Add identity to my existing MCP serverGovernMigration

Cookbook Conventions

All cookbooks follow the same structure:

  1. Goal — What you'll achieve
  2. Prerequisites — What you need before starting
  3. Time Estimate — How long it takes
  4. Steps — Numbered, actionable instructions
  5. Verify — How to confirm it's working
  6. Troubleshooting — Common issues and fixes
  7. Next Steps — Where to go from here

Code examples are copy-paste ready. SDK environment variables use the CHECKPOINT_ prefix; the Govern worker/deploy/self-host plane uses the separate AGENTSHIELD_ prefix for the same credentials — see Credentials for which one applies where.

Need Help?


Ready? Pick a cookbook above and let's build something.