Wind is a consumer wallet for AI apps. Users load one balance and spend it across every app they connect, so you can charge four cents for a background removal and still make money. It sits next to your Pro plan, not instead of it.
Users top up once. Every app they connect draws from the same balance, a few cents at a time.
Nobody pastes a provider key into your settings page. They tap Continue with Wind and they're in.
No $10 of credits stranded inside every app. One balance, spent wherever they are.
Charge for exactly what someone uses, down to a fraction of a cent, instead of selling their attention.
Pick a model and send a message. Wind holds an estimate, routes the call, meters the tokens that stream back, and charges only what was used.
Send a message to see it stream back.
Your margin settles to your account daily.
Wind does not replace your subscription. It sits underneath it and catches everyone who was never going to take it.
import { Wind } from '@wind/sdk';
const wind = new Wind({ apiKey });
await wind.charge(userId, { amount: 0.05 });
| Event | User | Charged | Wind | Net |
|---|---|---|---|---|
| thumbnail_pack | u_8f2a | $0.31 | $0.03 | $0.28 |
| bg_remove | u_31c9 | $0.04 | $0.00 | $0.04 |
| voice_clone | u_8f2a | $0.12 | $0.01 | $0.11 |
| export_hd | u_47bd | $0.25 | $0.03 | $0.22 |
Your users load a prepaid balance once. You charge them per action from it. The balance works across every app that accepts Wind, so it may already be funded before they reach you.
Two places you already have. The rate limit, where a free user hits the wall and won't buy Pro. And the cancel button, where a subscriber is leaving. Both currently lead to zero.
You control that. You set when Wind appears, and you set a monthly usage cap below your Pro price so the wallet can never be the cheaper path for a heavy user. Most people who take usage pricing weren't going to subscribe, but the cap means you don't have to take that on faith.
You can. Three things it doesn't solve. Your credits ask the user to prefund your app specifically, which is the same commitment decision they just declined. Whatever they don't spend is stranded with you. And you inherit fraud, chargebacks, refunds, tax and balance accounting. A Wind balance is already funded from somewhere else, so spending at your app costs the user nothing new.
Card fees. At 2.9% plus 30 cents, a 50-cent charge loses 63% to processing and anything under 31 cents costs more than it collects. That forces you to bundle up to $5 or more, and the bundle is the commitment your user already refused.
Nothing upfront, no minimum, nothing if nobody spends. We take a cut of what clears. The real cost is engineering time, and it's a payment flow to build and test like any other.
Yes. Per action, at whatever fits your product. We don't set it and we don't meter your tokens.
They happen at the load, not at your charges, so they sit with us rather than with you. That's a meaningful part of what you're buying.
Yes, and that's the honest friction. It's one signup and one card entry, after which every future purchase across every Wind app is a click.
Right now, most won't. That's why we're launching with a cluster of apps serving the same person rather than signing apps one at a time.
A balance you load once and spend across AI apps. Pay for what you actually use instead of a monthly subscription.
Because the same balance works everywhere Wind is accepted. If you only ever use one app, a subscription may suit you better. The wallet is for people using several tools occasionally.
Nothing. Wind is free to use. Apps pay us, not you.
There's a minimum because card fees make tiny loads uneconomical for us too. Whatever you load stays yours.
No.
Yes, on unspent balance.
No. You enter it once with Wind and never again at an individual app.
Depends on usage. If you use something daily, a subscription is probably better value. If you use it a few times a month, paying per use is almost always cheaper.
Load once. Spend anywhere Wind is connected. Pay for what you actually use.