Skip to main content
You can do anything with Narev pricing. Pick the path that matches how you ship.

​
Bare bones

Call the REST API. List providers, models, and prices. POST token counts and get a dollar cost back. No SDK required.

API reference

GET providers, models, and prices. POST /v1/traces/cost to price a call.

Quickstart

Copy-paste REST and MCP examples, then wire billing if you need it.

​
Billing

Price usage, then send dollar events to Polar, Stripe, Lago, OpenMeter, or another destination.
  • Any stack: price traces with the Narev API, then push events to your billing provider yourself.
  • Vercel AI SDK: wrap generateText / streamText with @ai-billing middleware. Cost resolution and destination routing happen in process.
Billing connections are self-managed in your application. Narev Cloud no longer hosts a Connect dashboard that stored provider credentials, provisioned meters, mapped models to products, or synced token prices into Stripe and Polar catalogs. Use @ai-billing destinations with environment variables, and rely on runtime cost resolution so catalog prices stay current without a sync job. For Stripe and Polar, you can alternatively run itemized per-model metering (separate meters per token type) and push price updates yourself using the Narev Prices API.

Node.js + Vercel AI SDK

Server middleware pattern for any Node runtime using @ai-billing.

Polar

Destination setup for Polar meters, customer sync, and event names.

Stripe

Destination setup for Stripe Meters, customer sync, and meter event payloads.

Lago

Destination setup for Lago billable metrics, customer sync, and event properties.

OpenMeter

Destination setup for OpenMeter CloudEvents, customer subjects, and meter types.

​
Warehouse

Land the live price catalog in Snowflake, BigQuery, Redshift, or whatever warehouse you already run. Use the Fivetran connector for a managed sync on a schedule.

Fivetran connector

Managed ELT. Sync Narev providers, models, and prices into your warehouse without writing a loader.