Vercel AI SDK Integration

Use the Vercel AI SDK to access 200+ models from multiple providers through a unified interface with the Martian Gateway.

Ensure you have your Martian API key from the Martian Dashboard before continuing.

Installation

Install the Martian provider and the Vercel AI SDK:

npm install @withmartian/ai-sdk-provider ai

If you already have the Vercel AI SDK installed, you only need to add @withmartian/ai-sdk-provider.

Configuration

Set your Martian API key in your .env file:

.env

MARTIAN_API_KEY=your_api_key_here

The martianProvider will read the API key from the MARTIAN_API_KEY environment variable.

Basic Usage

Quick Start

Use martianProvider with any AI SDK function to access models through the Martian Gateway:

import { generateText } from "ai";
import { martianProvider } from "@withmartian/ai-sdk-provider";

const { text } = await generateText({
  model: martianProvider("gpt-4o-mini"),
  prompt: "What is Olympus Mons?"
});

console.log(text);

The provider automatically prefixes OpenAI model names with openai/ for routing. gpt-4o-mini becomes openai/gpt-4o-mini.

Model Name Resolution

The Martian provider handles model naming:

OpenAI models: Automatically prefixed with openai/ if not already namespaced
Other providers: Already-namespaced models (e.g., anthropic/..., google/...) are passed through unchanged

import { generateText } from "ai";
import { martianProvider } from "@withmartian/ai-sdk-provider";

// OpenAI models - auto-prefixed
await generateText({
  model: martianProvider("gpt-4o-mini"), // → openai/gpt-4o-mini
  prompt: "Hello"
});

await generateText({
  model: martianProvider("openai/gpt-4o"), // → openai/gpt-4o
  prompt: "Hello"
});

// Any other model - passed through
await generateText({
  model: martianProvider("anthropic/claude-sonnet-4-20250514"),
  prompt: "Hello"
});

See the Available Models page for the complete list of supported models.

Chat Completions

Use the .chat() method for chat-based interactions:

import { generateText } from "ai";
import { martianProvider } from "@withmartian/ai-sdk-provider";

const { text } = await generateText({
  model: martianProvider.chat("gpt-4o"),
  messages: [
    { role: "user", content: "What is Olympus Mons?" }
  ]
});

console.log(text);

Text Completions

Use the .text() method for text completions:

import { generateText } from "ai";
import { martianProvider } from "@withmartian/ai-sdk-provider";

const { text } = await generateText({
  model: martianProvider.text("gpt-4o"),
  prompt: "Write a story about Mars. A long time ago... "
});

console.log(text);

Advanced Features

Streaming Responses

Use streamText for real-time streaming:

import { streamText } from "ai";
import { martianProvider } from "@withmartian/ai-sdk-provider";

const { textStream } = await streamText({
  model: martianProvider("gpt-4o"),
  prompt: "Write a story about Mars"
});

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

Tool Calling

Use the AI SDK's tool functionality with the Martian Gateway for function calling:

import { generateText } from "ai";
import { martianProvider } from "@withmartian/ai-sdk-provider";
import { z } from "zod";

const { text, toolCalls } = await generateText({
  model: martianProvider("anthropic/claude-sonnet-4-20250514"),
  prompt: "What's the weather in San Francisco?",
  tools: {
    getWeather: {
      description: "Get weather for a city",
      inputSchema: z.object({ city: z.string() }),
      execute: async ({ city }) => ({ city, temp: 60, conditions: "sunny" }),
    },
  },
  maxSteps: 3,
});

Structured Outputs

Generate structured data using schemas with generateObject:

import { generateObject } from "ai";
import { martianProvider } from "@withmartian/ai-sdk-provider";
import { z } from "zod";

const { object } = await generateObject({
  model: martianProvider("gpt-4o-mini"),
  schema: z.object({
    recipe: z.string(),
    ingredients: z.array(z.string()),
    steps: z.array(z.string()),
  }),
  prompt: "Generate a simple pasta recipe.",
});

console.log(object.recipe);
console.log(object.ingredients);

Advanced Generation

Use advanced parameters like system messages, multi-turn conversations, and generation settings:

import { generateText } from "ai";
import { martianProvider } from "@withmartian/ai-sdk-provider";

const { text, finishReason, usage } = await generateText({
  model: martianProvider("google/gemini-2.0-flash"),
  system: "You are a creative writing assistant.",
  messages: [
    { role: "user", content: "Write a short story about an alien." },
    { role: "assistant", content: "Once upon a time, there was an alien from Mars..." },
    { role: "user", content: "Continue the story with a twist ending." },
  ],
  temperature: 0.9,
  topP: 0.95,
  maxTokens: 800,
  stopSequences: ["THE END", "---"],
});

Next Steps

View Available Models

Browse 200+ AI models from leading providers with real-time pricing.

View Other Integrations

Explore other ways to integrate the Martian Gateway with your development workflow.

Code Examples

Practical examples and code snippets for common use cases.