Get started with Vibe in 2 minutes
Sign up, create an application, install the SDK, and make your first metered AI call. That's the whole setup.
Writing
UsageFlow Blog
Notes from the team on controlling how AI requests run: usage policies, runtime limits, and what a customer's balance can change.
Start here
Vibe Workflows: What Happens to an Identity's Usage
identity tells Vibe who's calling. A workflow tells it what to do about what they've used.
Get Started with Vibe: Your First Metered AI Call
Install it, wrap one call, and see the customer's credits get checked before the model runs.
How it works at runtime
Your App Needs a Usage Policy
Authentication says who someone is. It doesn't say how much they're allowed to do.
Enforce Usage Limits Where the Work Happens
A limit that runs after the work is a report. A limit that runs before it is a control.
AI Usage Isn't One Request = One Unit
A simple API knows every request costs 1. An AI request costs whatever it turns out to cost.
Stop AI Costs at Runtime: Block Calls at Zero
The cheapest overage is the one that never ran.
Billing & plans
Turn Usage Into a Customer Budget
A plan is just a budget with a name.
Why AI Apps Need a Credit System
A request is a bad unit of cost for AI. A credit is a good one.
Sell More AI Credits with Stripe
Stripe knows how to bill. It doesn't know what your AI product consumed. Vibe measures it and tells Stripe who it belongs to.