Agent Checkpoint

AI agents are already hitting your app: ChatGPT, Perplexity, Claude, and hundreds more are navigating your product on behalf of real users. Until now, that traffic was invisible.

Agent Checkpoint gives you real-time classification of every automated session across 40 signals: TLS ja4 fingerprint, user agent, path sequence, confidence score, and more. Every session is classified as ai_agent, bot, human, or incomplete_data.

Built for Next.js. Drop the middleware into your Vercel project and start seeing agent traffic in minutes.

Real-time classification with confidence scoring Per-agent history: pages visited, journey stage Webhook alerts when new agent types appear Full audit trail of every automated session

Protect your app with KYA-OS Checkpoint

Checkpoint detects AI-agent traffic to your app — and optionally blocks it on policy — with full visibility in the Checkpoint dashboard. Setup takes a few minutes.

1. Install the KYA-OS Checkpoint integration from the Vercel Marketplace and authorize it for your team.
2. Connect a project and pick a mode — Observe (classify and report, never block) or Enforce (block on policy). We provision a scoped CHECKPOINT_API_KEY and inject it into each selected project automatically.
3. Add the Checkpoint middleware so your app uses the key, then redeploy to pick up the new environment variable → Middleware installation guide.
4. Click Open in Checkpoint from the configuration screen to land authenticated in your project's dashboard and tune your policy.