How to Use Llama with the Vercel AI SDK — Meta Provider Guide

In this tutorial, you'll learn how to use Llama models with the Vercel AI SDK through Puter's OpenAI-compatible provider endpoint. No Meta API key needed — just your Puter auth token.

About Llama

Llama is Meta's large language model family, released under the Llama license. Being open-weight, Llama models can be self-hosted, fine-tuned, and run locally, though Puter lets you use them via API without managing infrastructure. Llama is popular for its strong general performance. Through Puter, you get Llama access via the Vercel AI SDK with no setup required.

Prerequisites

• A Puter account
• Your Puter auth token, go to puter.com/dashboard and click Copy to get your auth token

• Node.js installed on your machine

Setup

Install the Vercel AI SDK and the OpenAI provider:

npm install ai @ai-sdk/openai

Puter works as an OpenAI-compatible provider, so you use @ai-sdk/openai to connect. Configure it with Puter's base URL and your auth token:

import { createOpenAI } from '@ai-sdk/openai';

const puter = createOpenAI({
  baseURL: 'https://api.puter.com/puterai/openai/v1/',
  apiKey: 'YOUR_PUTER_AUTH_TOKEN',
});

Replace YOUR_PUTER_AUTH_TOKEN with the auth token you copied from your Puter dashboard. That's all you need. No Meta API key required.

Basic Text Generation

Here's a simple text generation call using Llama 4 Maverick:

import { createOpenAI } from '@ai-sdk/openai';
import { generateText } from 'ai';

const puter = createOpenAI({
  baseURL: 'https://api.puter.com/puterai/openai/v1/',
  apiKey: 'YOUR_PUTER_AUTH_TOKEN',
});

const { text } = await generateText({
  model: puter.chat('meta-llama/llama-4-maverick'),
  prompt: 'What is the capital of France?',
});

console.log(text);

The code is identical to what you'd write for any OpenAI provider. The only difference is the base URL and the model string.

Streaming

For longer responses, use streamText to get results in real-time:

import { createOpenAI } from '@ai-sdk/openai';
import { streamText } from 'ai';

const puter = createOpenAI({
  baseURL: 'https://api.puter.com/puterai/openai/v1/',
  apiKey: 'YOUR_PUTER_AUTH_TOKEN',
});

const result = streamText({
  model: puter.chat('meta-llama/llama-4-maverick'),
  prompt: 'Write a short story about a robot learning to paint.',
});

for await (const chunk of result.textStream) {
  process.stdout.write(chunk);
}

Use streamText instead of generateText and iterate over result.textStream to get text chunks as they arrive.

Why Use Puter?

You could use Llama through a hosting provider's API directly. Here's why Puter is a simpler option:

• One API key for everything — no need to sign up for separate hosting providers, Anthropic, or OpenAI accounts. Your Puter auth token covers all providers.
• One setup for all models — the same Puter config works for Claude, GPT, Gemini, Llama, and 400+ other models. Just change the model string.
• No extra packages — without Puter, each AI provider needs its own SDK package and API key. With Puter, everything goes through a single @ai-sdk/openai setup.

Conclusion

You now have the Meta provider set up through the Vercel AI SDK via Puter — no API key needed. Swap the model string to use any Llama model, from the lightweight Llama 3 to the powerful Llama 4 Maverick, or any of the hundreds of other AI models available through Puter.

Related

• How to Use Vercel AI SDK with Puter
• Free, Unlimited Llama API
• Meta Llama Models on Puter
• Access DeepSeek Using Vercel AI SDK
• Access Mistral Using Vercel AI SDK
• Supported AI Models