Usage Based Billing: for You and Your Agents.

Simple enough to power your grandma's (👵) AI B2B SaaS.

Three ways to bill

All of them are terrible ideas. But we help simplify ALL of your horrendous decisions. No questions asked.

Billing the Vercel AI SDK hasn't been easier.

One import. One call. Your AI SDK calls are now billed automatically. No webhooks, no spreadsheets, no crying in the bathroom.

import { generateText } from 'ai';
import { scrawn } from 'scrawn/vercel';

const { text } = await generateText({
  model: scrawn('gpt-4'),
  prompt: 'I want you to bill me like one of your French girls',
});

import { generateText } from 'ai';import { scrawn } from 'scrawn/vercel';const { text } = await generateText({  model: scrawn('gpt-4'),  prompt: 'I want you to bill me like one of your French girls',});

const usage = await db.logs.count({
  where: { userId },
});
if (usage % 100 === 0) {
  await stripe.invoice.create({ ... });
  await slack.send('#billing-prayers');
}
// TODO: reconcile (lol)

const usage = await db.logs.count({  where: { userId },});if (usage % 100 === 0) {  await stripe.invoice.create({ ... });  await slack.send('#billing-prayers');}// TODO: reconcile (lol)

Bill your self rolled abomination in one-ish line.

You have a Stripe integration, three cron jobs, a Google Sheet, and a prayer. The dishwasher is right there, but you choose the pain. Respect. We did what we could.

Charge per API request

No metering. No tracking. Just vibes. Your investors will either applaud your hustle or file a lawsuit. There is no in-between.

export async function POST(req) {
  await charge(9.99); // per message?
  return streamText({
    model: openai('gpt-4'),
    prompt: 'u up?',
  });
}

export async function POST(req) {  await charge(9.99); // per message?  return streamText({    model: openai('gpt-4'),    prompt: 'u up?',  });}