Protect margin
Keep expensive usage inside policy.
Check allowances, limits, and overage rules before an AI call, agent action, or API request creates cost.
Usage infrastructure for AI and API products
Turn variable AI usage into predictable revenue.
Meter every valuable action, enforce plan limits before expensive work starts, forecast customer runout, and send trusted usage to Stripe or the commerce systems you already use.
5,000 calls/month includedTypeScript SDK or HTTPStripe synchronization available
Used by
Connects withstripeOpenTelemetrySlack
Three commercial outcomes
Control usage economics without rebuilding your product.
Give customers clarity
Show usage before the invoice.
Embed current consumption, allowance, forecast, and the next useful action directly inside your product.
Act before runout
Turn forecasts into timely decisions.
See which customers are approaching a limit, behaving unusually, or ready for a larger package while there is time to respond.
One operating loop
Meter → Control → Forecast → Synchronize.
Instrument once, then change commercial policy as pricing and customer behavior evolve.
Meter
Capture the product events that create value and variable cost.
Control
Apply customer allowances, limits, throttles, and overage policy.
Forecast
Predict runout, changes in behavior, and expansion signals.
Synchronize
Send validated usage to Stripe or the systems you already use.
One reversible wrapper, every existing call
Wrap the client once.
Keep the OpenAI, Anthropic, or OpenRouter calls you already ship. UsageTap wraps the client to meter outcomes, apply policy, and feed forecasts and billing. Remove one wrapper to roll back.
Read the full integration guide →server.tsUsageTap SDK
import OpenAI from "openai";
import { withMetering } from "@usagetap/sdk/openai";
const openai = withMetering(
new OpenAI(),
"cust_123",
);
// Existing OpenAI calls stay the same.
const response = await openai.responses.create({
model: "gpt-5.5-mini",
input: "Generate the document",
});Design before you deploy
Turn a pricing idea into allowances, overages, and margin.
Model credits, actions, workflows, or hybrid usage pricing with your own costs and customer assumptions. Then use the result to guide your Meter setup.
A useful first decision
What should the customer pay for?
AI actionWorkflowGenerated outputAPI callCreditHybrid plan
A separate product for input cost
Need to cut the tokens inside each request?
UsageTap Compress protects critical text, removes low-value context, and gives you an inspectable result to evaluate before production rollout.
Start with real usage
Make AI products commercially predictable.
Connect your first customer and meter, then add controls, forecasts, customer visibility, and commerce sync as you grow.