MONETIZING AGENT TRAFFIC
How do you monetize AI agent traffic?
Last reviewed: 2026-10-10. The paid-link explainer walks through the same mechanism request by request.
By pricing the request instead of the impression. The resource answers an automated request with an HTTP 402 offer naming the amount, the token, the network and the receiving address, and the content is released only after that payment settles. A client that does not pay is left with the offer and no content.
The price travels with the resource, so a person, a crawler and an agent are quoted the same amount, and that amount is visible before anything is authorised.
That is what separates this from advertising and affiliate revenue: an ad needs a rendered page and human attention, and an affiliate pays only when a sale completes, so neither is triggered by a machine that reads the content and leaves.
Why an unpaid visit stays unpaid
An impression needs a person
Advertising is billed on a rendered view and human attention, and an agent produces neither.
A crawler that fetches a page, reads the markup and leaves creates no view for an advertiser to buy. The visit still costs bandwidth and origin work, and the ad revenue it would have produced stays at zero.
An affiliate link needs a completed sale
Affiliate programmes pay on a qualifying purchase, which an agent reading a page to answer a question rarely makes.
The click can be logged and a cookie can be set, but the payout arrives only for the sale the programme counts. A read that never converts is an unpaid visit, whether a person or a machine made it.
Blocking a name is not pricing it
A user-agent string is a claim, not a payment term, and the same content served under a different name still earns nothing.
Charging needs a term the other side can meet. A price on the resource is that term: every client, whatever name it sends, is quoted the same amount and can accept it.
What pricing a request actually means
The price belongs to the resource
Protected content is not printed into a page. It answers at its own endpoint with an HTTP 402 offer: the amount, the token, the network and the receiving address, and no part of the content.
Payment settles before delivery
The client returns a signed authorisation for that exact request. The payment is verified and the resource released only after settlement, and a redirect, a screenshot or a browser flag is never treated as proof of payment.
One transfer per request
Settlement is a USDC transfer on the configured network, publicly recorded and inspectable by the payer. There is no advertising network in the middle and no impression to count.
No crawler name required
Access depends on a verified payment rather than on the identity a visitor claims. A visitor type (AI or human) is recorded here for reporting only, and it never grants or denies access.
Availability of each route
New token payments are paused while the wallet-security review is unresolved. The price, network and receiving address stay configured, and they are not an invitation to pay.
Card checkout is open at US$19.00 USD, charged once through Stripe, with no wallet required.
Ownership of the original wallet and the original transaction are checked, and recovery never asks for a second payment.
The report area shows five synthetic views on fictional data. New report purchases are on hold.
Questions and short answers
Is AI agent traffic worth anything?
It is worth something when a client can and will pay for the resource it asks for. That is the entire condition: nothing on this site promises a number of paying visits, a conversion rate or a monthly figure.
Do I have to detect the agent before I charge it?
No. The price is attached to the resource, so any client can accept it, and no crawler name is needed for a payment to be made.
Does charging for requests conflict with robots.txt?
They are separate conditions. robots.txt states which paths may be crawled, and this site's paid endpoint is excluded from crawling. Paying and being permitted to crawl are different questions.
Can my own link be charged this way today?
Not self-serve today: the paid route serves PayPerAI's own resources. The API guide documents the mechanism, and the contact page is the way to raise a question about your own link.
What does a request cost here?
The price is set per resource and shown to the requester before anything is authorised. Each offer names the exact amount, the token, the network and the receiving address.
Do human visitors pay to read this site?
No. The storefront stays free to read, and only a protected resource endpoint asks for payment.
Terms this page relies on
- HTTP 402 Payment Required
- The HTTP status code reserved for payment conditions. A server returns it to say that the requested resource is available for a price rather than for free.
- Monetized link
- A link whose response depends on a payment condition: the visitor is quoted a price, and the content is released only after that payment is verified.
- Per-request pricing
- Pricing a resource by the request instead of by an advertisement impression. The price travels with the resource, so every requester sees the same amount.
- Settlement
- The point at which a payment is recorded on the network and the funds become the recipient's. A signed authorisation on its own is not settlement.
- User-agent
- The text a client sends to identify itself. It can be copied by anyone, so it is treated as reporting data and never as identity or authorisation.
- x402
- An open convention for charging for a request over HTTP: the server answers with the price and payment details, the client returns a signed authorisation, and the resource is released once the payment settles.
All definitions lists the full set of terms in one place.
Where the practical material is
The paid route is the layer that charges for a resource. The Ecommerce Supplier & Margin Pack v1 is the download that is actually on sale today: a reusable starting point for checking suppliers, organising catalogue data and calculating order economics.