Under the hood
How Marech Works
Marech works in two modes. Paste a one-line snippet to monitor traffic anywhere, and add an edge integration (Cloudflare Worker, WordPress plugin, or proxy) to actually block scrapers before your content is served. Here's the flow for the blocking path.
1
Request hits the edge
The Cloudflare Worker, WordPress plugin, or proxy sees the request before your origin serves any HTML.
2
Is it an AI bot?
Marech checks user agent, headers, IP reputation, and behavior patterns in real time.
3
Decision: Block or Allow
Based on your policies, the bot gets a 403 and never sees your content — or is allowed through. (If Marech is unreachable, it fails open and serves the page.)
4
Logged in your dashboard
Every decision is recorded — bot type, IP, path, confidence score, and action taken.
How we detect AI bots
01
Known Bot Database
We maintain a constantly-updated list of 20+ known scrapers — GPTBot, ClaudeBot, CCBot, Amazonbot, PerplexityBot, and more.
02
Behavior Analysis
Detect headless browsers and bots pretending to be humans by analyzing missing headers, timing patterns, and JavaScript execution.
03
Datacenter IP Detection
Real users browse from home and mobile networks. AI scrapers come from cloud datacenters. We flag datacenter IPs automatically.
You're in control
Policies let you decide exactly what gets blocked. No one-size-fits-all rules.
Block all AI scrapers
The default. Zero AI bots get through — GPTBot, Claude, and 20+ others.
Block during business hours only
Let archive bots in at night while protecting during peak traffic.
Block on specific paths
Protect /admin, /api, or /checkout while leaving /blog open.