This is the official provider SDK for 302AI based on the v6 (Beta) version, integrated with the Vercel AI SDK.

The 302AI provider enables seamless integration with the Vercel AI SDK, offering access to a wide range of AI models for:

• Advanced language models including:
  • GPT 5.1 series (gpt-5.1-plus, gpt-5.1-codex, etc.)
  • Claude 4-5 series (opus, sonnet, haiku)
  • Gemini series
  • Llama 3 series
  • And many other leading models

• State-of-the-art image models including:
  • Midjourney 6.0/6.1/7.0
  • GPT-Image-1
  • Stable Diffusion variants (SD3, SDXL)
  • Flux models
  • Ideogram models
  • Qwen models
  • Doubao models (Seedream 3.0/4.0)
  • Gemini-2.5-flash-image (Nano-banana)、Gemini-3-pro-image-preview (Nano-banana-pro)
  • And more specialized image generators

• High-quality embedding models including:
  • OpenAI text-embedding-3 (small/large)
  • BGE models
  • Zhipu embeddings
  • And other specialized embedding models

• Text-to-Speech with 10+ providers:
  • OpenAI (alloy, nova, shimmer, etc.)
  • Azure Cognitive Services
  • ElevenLabs
  • Doubao (ByteDance)
  • Google Cloud TTS
  • Qwen (Alibaba)
  • And more specialized TTS providers
• Features:
  • ⚡ Sync and async modes
  • 🎚️ Customizable speed, volume, emotion
  • 📦 Multiple output formats (mp3, wav, ogg)

All models are accessible through a unified, type-safe API that follows Vercel AI SDK standards. For a complete list of supported models, please refer to our detailed documentation.

• 🚀 Simple, unified API for all AI operations
• 💪 Full TypeScript support
• 🔄 Streaming responses support
• 🛡️ Built-in error handling
• 📝 Comprehensive documentation

npm install @302ai/ai-sdk

import { ai302 } from '@302ai/ai-sdk';
import { generateText } from 'ai';

// Generate text using GPT-4 Turbo
const { text } = await generateText({
  model: ai302('gpt-4-turbo'),
  prompt: 'Hello, how can I help you today?'
});

console.log(text);

import { experimental_generateImage as generateImage } from 'ai';

// Generate image with Midjourney
const { image } = await generateImage({
  model: ai302.image('midjourney/6.1'),
  prompt: 'A beautiful sunset over mountains'
});

import { embed } from 'ai';

// Generate text embeddings
const { embedding } = await embed({
  model: ai302.textEmbeddingModel('text-embedding-3-large'),
  value: 'Text to embed'
});

console.log(embedding);

import { generateSpeech } from 'ai';
import fs from 'fs';

// Generate speech file
const { audio } = await generateSpeech({
  model: ai302.speech('openai/alloy'),
  text: 'Hello, welcome to 302AI!'
});

fs.writeFileSync('speech.mp3', audio.uint8Array);

This SDK includes the following packages:

💡 Each package is 100% TypeScript and provides full type support.

For detailed documentation about available models and usage, please check:

• 📘 302AI Provider Documentation
• 📗 Vercel AI SDK Documentation

Apache License 2.0 - See LICENSE for details.