> ## Documentation Index
> Fetch the complete documentation index at: https://narev.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Getting started

> First steps with Fireworks middleware for tracking AI costs.

## Installation

<CodeGroup>
  ```bash npm theme={null}
  npm install @ai-billing/fireworks @ai-billing/core @ai-sdk/fireworks ai
  ```

  ```bash pnpm theme={null}
  pnpm add @ai-billing/fireworks @ai-billing/core @ai-sdk/fireworks ai
  ```

  ```bash yarn theme={null}
  yarn add @ai-billing/fireworks @ai-billing/core @ai-sdk/fireworks ai
  ```
</CodeGroup>

## Overview

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

It captures Fireworks' OpenAI-compatible usage payload, including cached prompt tokens and reasoning tokens, ensuring that costs are accurately reflected.

## Usage

To use the middleware, wrap your Fireworks model using `wrapLanguageModel` from the `ai` package and pass the `createFireworksMiddleware`.

<Steps>
  <Step title="Initialize the Fireworks provider">
    First, set up the provider using the Fireworks SDK and your API key.

    ```typescript theme={null}
    import { createFireworks } from '@ai-sdk/fireworks';

    const fireworks = createFireworks({
      apiKey: process.env.FIREWORKS_API_KEY,
    });
    ```
  </Step>

  <Step title="Define model pricing">
    Set up a price resolver to define the costs for the models you'll be using. Fireworks bills on
    prompt tokens, completion tokens, and cached prompt tokens (`inputCacheReadTokens`).

    Fireworks has no separate pricing tier for reasoning tokens — reasoning tokens are billed at the
    same rate as `completionTokens`, so you don't need to configure anything extra for reasoning
    models.

    ```typescript theme={null}
    import { createObjectPriceResolver } from '@ai-billing/core';

    const priceResolver = createObjectPriceResolver({
      'accounts/fireworks/models/glm-5p3-flash': {
        promptTokens: 0.0000002, // 0.2 per 1M tokens
        completionTokens: 0.0000008, // 0.8 per 1M tokens
        inputCacheReadTokens: 0.0000001, // 0.1 per 1M tokens
      },
    });
    ```
  </Step>

  <Step title="Create the billing middleware">
    Initialize the Fireworks billing middleware. You need to provide a destination (such as `consoleDestination`) where billing events will be sent, along with your `priceResolver`.

    ```typescript theme={null}
    import { createFireworksMiddleware } from '@ai-billing/fireworks';
    import { consoleDestination } from '@ai-billing/core';

    const billingMiddleware = createFireworksMiddleware({
      destinations: [consoleDestination()],
      priceResolver: priceResolver,
    });
    ```
  </Step>

  <Step title="Wrap the model">
    Use `wrapLanguageModel` from the `ai` package to apply the billing middleware to your Fireworks model.

    ```typescript theme={null}
    import { wrapLanguageModel } from 'ai';

    const wrappedModel = wrapLanguageModel({
      model: fireworks('accounts/fireworks/models/glm-5p3-flash'),
      middleware: billingMiddleware,
    });
    ```
  </Step>

  <Step title="Use the wrapped model">
    Finally, use the wrapped model with AI SDK functions like `generateText` or `streamText`. The billing middleware will automatically track tokens, handle reasoning metrics, and calculate costs.

    ```typescript theme={null}
    import { generateText } from 'ai';

    const result = await generateText({
      model: wrappedModel,
      prompt: 'What is the capital of Sweden?',
    });
    ```
  </Step>
</Steps>