Getting started

Installation

npm install @ai-billing/cerebras @ai-billing/core @ai-sdk/cerebras ai

pnpm add @ai-billing/cerebras @ai-billing/core @ai-sdk/cerebras ai

yarn add @ai-billing/cerebras @ai-billing/core @ai-sdk/cerebras ai

Overview

The @ai-billing/cerebras package provides middleware for tracking token usage and calculating costs when using Cerebras models with the Vercel AI SDK.

Usage

To use the middleware, wrap your Cerebras model using wrapLanguageModel from the ai package and pass the createCerebrasMiddleware.

Initialize the Cerebras provider

First, set up the Cerebras provider using your API key.

import { createCerebras } from '@ai-sdk/cerebras';

const cerebras = createCerebras({
  apiKey: process.env.CEREBRAS_API_KEY,
});

Define model pricing

Set up a price resolver to define the costs for the models you’ll be using. You can use either the dynamic Narev price resolver or a manual object price resolver.

Using Narev Pricing (Recommended)
Using Manual Pricing

The Narev price resolver automatically fetches the latest pricing for Cerebras models.

import { createNarevPriceResolver } from '@ai-billing/core';

const priceResolver = createNarevPriceResolver({
  apiKey: process.env.NAREV_API_KEY ?? '',
});

You can also manually specify costs for standard prompt and completion tokens.

import { createObjectPriceResolver } from '@ai-billing/core';

// Pricing per Cerebras's own docs at the time of writing:
// https://inference-docs.cerebras.ai/models/openai-oss ($0.35 / $0.75 per 1M tokens).
// Cerebras pricing can change — verify against the link above before relying on this in
// production.
const priceResolver = createObjectPriceResolver({
  'gpt-oss-120b': {
    promptTokens: 0.35 / 1_000_000,
    completionTokens: 0.75 / 1_000_000,
  },
});

Create the billing middleware

Initialize the Cerebras billing middleware. You need to provide a destination (such as consoleDestination) where billing events will be sent, along with your priceResolver.

import { createCerebrasMiddleware } from '@ai-billing/cerebras';
import { consoleDestination } from '@ai-billing/core';

const billingMiddleware = createCerebrasMiddleware({
  destinations: [consoleDestination()],
  priceResolver: priceResolver,
});

Wrap the model

Use wrapLanguageModel from the ai package to apply the billing middleware to your Cerebras model.

import { wrapLanguageModel } from 'ai';

const wrappedModel = wrapLanguageModel({
  model: cerebras('gpt-oss-120b'),
  middleware: billingMiddleware,
});

Use the wrapped model

Finally, use the wrapped model with AI SDK functions like generateText or streamText. The billing middleware will automatically track tokens and calculate costs.

import { generateText } from 'ai';

const result = await generateText({
  model: wrappedModel,
  prompt: 'What is the capital of Sweden?',
});

Overview

@ai-billing SDK

narevai/thin-ops

Installation

Overview

Usage

​Installation

​Overview

​Usage

Installation

Overview

Usage