Feature
Detect AI agents and classify every request into a lane.
Wayleave sorts every request into one of four lanes before your route runs, so you can treat an AI assistant, an undisclosed scraper and a person differently — and log the evidence behind each decision.
Four lanes for every request
verified_agentA valid Ed25519 HTTP Message Signature (RFC 9421, compatible with Web Bot Auth) against a key directory you trust.
declared_agentIdentifies itself as a bot or AI agent, but carries no valid signature.
suspected_botAutomation fingerprints without disclosure. An invalid signature lands here too — faking verification is the strongest fraud signal there is.
humanBrowser-shaped traffic. Nothing gave it away — humans are never charged.
Detection you can explain
Every decision returns its evidence and is loggable. Only the verified lane is cryptographic; the other three are read from what the client says about itself.
Know the limits
Wayleave prices disclosure; it does not detect concealment. A bot that sends a browser user-agent and accept-language header reaches the human lane — set strictPricedPaths: true on anything you charge for, so only a verified signature (or your own confirmHuman check) crosses a priced route.
Observe AI agent traffic
npm install wayleave
import Wayleave from 'wayleave';
const gate = new Wayleave({
meter: { apiKey: process.env.WAYLEAVE_METER_KEY },
});
app.use(gate.express()); // before the routes you want to observeQuestions
Does Wayleave block all bots?
No — it is a screening and pricing layer, not a wall. Pair it with your edge or CDN for concealment detection such as TLS fingerprinting.