● Early pilot — advertisers welcome

Get recommended by AI agents — honestly.

Fairplug lets agents surface your offer clearly labelled as sponsored, with a tracked link. You pay only when a sale is validated. Nothing during the pilot.

The problem

People now ask AI agents what to buy

AI agents increasingly recommend products, and people act on those answers. But if you sell something good, there is no honest way to be recommended. You can't tell an agent about your offer, and nobody can see why one product was picked over another.

The fix

A clear, labelled offer

Fairplug gives agents a standard way to find your offer. When an agent recommends it, the answer says it is sponsored and uses a tracked link, so sales can be counted and validated. Nothing is hidden from the person asking.

How it works

Three steps

Kept simple on purpose while we are early.

1

You list your offer

Tell us what you sell, where you ship, and the commission or payout you can offer per sale.

2

Agents surface it

Agents can recommend it when it fits, clearly labelled as sponsored, with a tracked link.

3

You pay on validated sales

Only sales that are checked and confirmed count. No sale, no fee.

Pricing

Open book, pay only on validated sales

The fee is published and the same for everyone.

10%

of validated commissions or agreed payout

  • US$0.50 minimum per validated sale
  • 15% cap — we never take more
  • Charged only on validated sales
  • No payment during the pilot

What to expect

Fairplug is early and unproven. We have no track record of sales yet, and we can't promise volume or results. The pilot is a chance to try it at no cost and help shape it.

Sign up

List your product

No account or payment needed. The button opens an email with the details we need already laid out. Fill them in and send.

Your email will ask for:
Company:
Product URL:
What you sell:
Commission or payout you can offer:
Region you ship to:
Open prefilled email →

Button not working? Email fairplughq@gmail.com with those five details.

For builders. Building an agent instead? Try the sandbox MCP endpoint or read the source on GitHub.