OpenRouter Provider for Vercel AI SDK

The OpenRouter provider for the Vercel AI SDK gives access to over 300 large language model on the OpenRouter chat and completion APIs.

Setup

1# For pnpm
2pnpm add @openrouter/ai-sdk-provider
3
4# For npm
5npm install @openrouter/ai-sdk-provider
6
7# For yarn
8yarn add @openrouter/ai-sdk-provider

Provider Instance

You can import the default provider instance openrouter from @openrouter/ai-sdk-provider:

1import { openrouter } from '@openrouter/ai-sdk-provider';

Example

1import { openrouter } from '@openrouter/ai-sdk-provider';
2import { generateText } from 'ai';
3
4const { text } = await generateText({
5  model: openrouter('openai/gpt-4o'),
6  prompt: 'Write a vegetarian lasagna recipe for 4 people.',
7});

Supported models

This list is not a definitive list of models supported by OpenRouter, as it constantly changes as we add new models (and deprecate old ones) to our system.
You can find the latest list of models supported by OpenRouter here.

You can find the latest list of tool-supported models supported by OpenRouter here. (Note: This list may contain models that are not compatible with the AI SDK.)

Passing Extra Body to OpenRouter

There are 3 ways to pass extra body to OpenRouter:

1. Via the providerOptions.openrouter property:

   1import { createOpenRouter } from '@openrouter/ai-sdk-provider';
   2import { streamText } from 'ai';
   3
   4const openrouter = createOpenRouter({ apiKey: 'your-api-key' });
   5const model = openrouter('anthropic/claude-3.7-sonnet:thinking');
   6await streamText({
   7  model,
   8  messages: [{ role: 'user', content: 'Hello' }],
   9  providerOptions: {
   10    openrouter: {
   11      reasoning: {
   12        max_tokens: 10,
   13      },
   14    },
   15  },
   16});

2. Via the extraBody property in the model settings:

   1import { createOpenRouter } from '@openrouter/ai-sdk-provider';
   2import { streamText } from 'ai';
   3
   4const openrouter = createOpenRouter({ apiKey: 'your-api-key' });
   5const model = openrouter('anthropic/claude-3.7-sonnet:thinking', {
   6  extraBody: {
   7    reasoning: {
   8      max_tokens: 10,
   9    },
   10  },
   11});
   12await streamText({
   13  model,
   14  messages: [{ role: 'user', content: 'Hello' }],
   15});

3. Via the extraBody property in the model factory.

   1import { createOpenRouter } from '@openrouter/ai-sdk-provider';
   2import { streamText } from 'ai';
   3
   4const openrouter = createOpenRouter({
   5  apiKey: 'your-api-key',
   6  extraBody: {
   7    reasoning: {
   8      max_tokens: 10,
   9    },
   10  },
   11});
   12const model = openrouter('anthropic/claude-3.7-sonnet:thinking');
   13await streamText({
   14  model,
   15  messages: [{ role: 'user', content: 'Hello' }],
   16});

Anthropic Prompt Caching

You can include Anthropic-specific options directly in your messages when using functions like streamText. The OpenRouter provider will automatically convert these messages to the correct format internally.

Basic Usage

1import { createOpenRouter } from '@openrouter/ai-sdk-provider';
2import { streamText } from 'ai';
3
4const openrouter = createOpenRouter({ apiKey: 'your-api-key' });
5const model = openrouter('anthropic/<supported-caching-model>');
6
7await streamText({
8  model,
9  messages: [
10    {
11      role: 'system',
12      content:
13        'You are a podcast summary assistant. You are detail oriented and critical about the content.',
14    },
15    {
16      role: 'user',
17      content: [
18        {
19          type: 'text',
20          text: 'Given the text body below:',
21        },
22        {
23          type: 'text',
24          text: `<LARGE BODY OF TEXT>`,
25          providerOptions: {
26            openrouter: {
27              cacheControl: { type: 'ephemeral' },
28            },
29          },
30        },
31        {
32          type: 'text',
33          text: 'Who're the speakers?',
34        },
35      ],
36    },
37  ],
38});