Let the crawlers in,
challenge the rest.
Classify every request — verified crawler, automated client or likely human — and allow, challenge or block it. Verified bots like Googlebot always pass, so your SEO never breaks. The flat-rate alternative to Cloudflare Bot Management and Vercel BotID.
Scrapers and abuse shouldn’t cost you Googlebot
The hard part of blocking bots is not blocking the good ones. koigrid classifies the user-agent into verified crawler (Googlebot, Bingbot, GPTBot…), automated client (curl, python-requests, headless Chrome, scanners) or likely human, and your edge calls one endpoint to get an allow / challenge / block decision. Verified crawlers are allowlisted by default so you can’t deindex yourself. Set the mode to challenge and suspicious clients get a proof-of-work Attack Challenge — a signed puzzle they must solve before you grant clearance — no CAPTCHAs, no third-party edge.
How koigrid compares
| koigrid | Cloudflare Bot Management | |
|---|---|---|
| Pricing | Flat rate | Enterprise / per request |
| Verified-bot allowlist | Yes (default) | Yes |
| Automated vs human | Yes | Yes (ML score) |
| Proof-of-work challenge | Yes | Yes (managed) |
| Decision API for your edge | Yes | Runs at CF edge |
| Manage by AI agent | Native | Dashboard/API |
From public docs, 2026. v1 is a classifier + decision endpoint + proof-of-work challenge your edge enforces.
Bot control that respects SEO
SEO-safe
Googlebot, Bingbot, GPTBot and friends are allowlisted by default — never block them by accident.
Three-way classify
Verified crawler, automated client or likely human — from the user-agent.
Allow / challenge / block
Pick a mode per project; your edge calls one endpoint per request.
Proof-of-work challenge
A signed puzzle suspicious clients must solve — no CAPTCHA, no third party.
Bring your own edge
A decision you enforce anywhere — not locked to one edge network.
Agent-native
Set the policy from Claude Code or any LLM via API + CLI.