People bring people.
You just count them right.
Links, attribution, and email invites over a two-call API your coding agent integrates in a single pass. Idempotent, so every human is counted once.
You’re on the list.We’ll send one email the moment the private beta opens, nothing else.
Small first cohort · or read llms.txt first
POST /v1/signups
Small on purpose.
Two endpoints. One key. The surface is small enough that an agent reads it once and gets the integration right, and small enough that you can audit what it does in an afternoon.
Retries are the normal case in agent-written code, not the exception. That is why the idempotency key is part of the contract, not an afterthought.
We own the loop up to money.
The referral loop, end to end.
- Mint links and codes, one per referrer, ready to share anywhere.
- Capture signups, bind a new account to the referrer whose code it carried.
- Invite by email, the referrer submits addresses and we send from shared infrastructure. Zero setup.
- Protect deliverability, syntax and MX validation plus rate limits before anything sends.
- Report attributions, who referred whom, counts per referrer, over a clean API.
Deliberately out of scope.
- Reward payouts, money-touching and app-specific. We report the attribution, you decide what it is worth.
- Owning your users, no re-verification. We trust your “this user is real” signal and throttle.
- A dashboard you live in, the API is the surface. A light reporting UI comes later, for humans who want a look.
Same referral links. Different reader.
Every product below does referrals well for a marketing team clicking through a dashboard. The difference is what happens when the thing doing the integration is a coding agent.
Built for a human
A dashboard, a JavaScript snippet, and a support article when the snippet does not fire. The integration assumes a person is watching.
Built for an agent
A machine-readable contract an agent can integrate and verify without a human in the loop. Self-explaining errors, and working referrals on the first try.
From “I need referrals” to captured, in one sitting.
Point your agent at what we publish.
It reads llms.txt for the shape of the product, and the scaffold generator prints the exact calls for your framework. No SDK archaeology.
Mint a code, capture the signup.
One call issues a referrer’s code. One call attributes the account that used it.
Watch your best users compound.
Every attributed signup tells you which customers actually bring you more customers, so you can put real budget behind the ones who do.
The short answers.
What is simplereferral?
simplereferral is a referral API designed for AI coding agents to integrate correctly on the first try. It mints referral links, captures signup attribution, and sends email invites using one API key and two idempotent calls.
What happens if my agent retries the same call?
It replays. The count does not move.
How is simplereferral different from Rewardful, FirstPromoter, GrowSurf, or Tolt?
Those tools are built for a human configuring a dashboard. simplereferral is built for an AI coding agent writing integration code: a tiny predictable surface, idempotent endpoints, actionable errors, and docs designed to be machine-readable, with an llms.txt today and an OpenAPI spec and MCP server arriving with the beta.
Does simplereferral handle reward payouts?
No. simplereferral reports attributions, who referred whom, and deliberately leaves reward payouts to your app, since payouts are money-touching and app-specific.
What does it cost?
Pricing is not set yet. First-cohort accounts keep the terms they join on.
Can I see it work before I give you an email?
Yes. llms.txt is public, and the sandbox runs a real capture against a demo tenant, keyless, in your browser.
Small first cohort. Founding terms.
You get a key, a direct line to the founder, and terms that stay put.
You’re on the list.Watch for one email when the private beta opens.
Small first cohort · or read llms.txt first