@wix/ai

Wix AI SDK for the Vercel AI SDK. Provides access to AI models through Wix infrastructure with automatic authentication.

⚠️ Note: This package depends on Vercel AI SDK v6 (currently in beta).

Supported Providers

| Provider | Status | Capabilities | |----------|--------|--------------| | OpenAI | ✅ Available | Text generation, chat, embeddings | | Runware | ✅ Available | Image generation |

Installation

npm install @wix/ai

Requirements:

• ai@>=6.0.27 (Vercel AI SDK v6)
• zod@^4.1.8

Quick Start

Wix Environment (Blocks, Wix CLI Apps)

import { generateText } from 'ai';
import { openai } from '@wix/ai';

const { text } = await generateText({
  model: openai('gpt-4o'),
  prompt: 'What is the capital of France?',
});

Self-Hosted Apps

For apps running outside Wix, pass a WixClient configured with AppStrategy or ApiKeyStrategy:

import { generateText } from 'ai';
import { createOpenAI } from '@wix/ai';
import { createClient, AppStrategy } from '@wix/sdk';

const wixClient = createClient({
  auth: AppStrategy({
    appId: 'YOUR_APP_ID',
    appSecret: 'YOUR_APP_SECRET',
    instanceId: 'YOUR_INSTANCE_ID',
  }),
});

const openai = createOpenAI({ client: wixClient });

const { text } = await generateText({
  model: openai('gpt-4o'),
  prompt: 'What is the capital of France?',
});

Requirements

• ✅ Wix backend code (API routes, server functions)
• ✅ Self-hosted apps (with WixClient using AppStrategy or ApiKeyStrategy)
• ❌ Browser/client-side code

Technical requirements:

• Node.js 18+
• ai package v6.0.0+

API Reference

OpenAI Provider

Responses API (Default)

import { generateText } from 'ai';
import { openai } from '@wix/ai';

const { text } = await generateText({
  model: openai('gpt-4o'),
  prompt: 'Explain quantum computing.',
});

Or explicitly:

const model = openai.responses('gpt-4o');

Chat Completions

import { generateText } from 'ai';
import { openai } from '@wix/ai';

const { text } = await generateText({
  model: openai.chat('gpt-4o'),
  messages: [
    { role: 'system', content: 'You are a helpful assistant.' },
    { role: 'user', content: 'What is TypeScript?' },
  ],
});

Embeddings

import { embed, cosineSimilarity } from 'ai';
import { openai } from '@wix/ai';

const { embedding } = await embed({
  model: openai.embeddingModel('text-embedding-3-small'),
  value: 'Hello world',
});

const similarity = cosineSimilarity(embedding1, embedding2);

Batch embeddings:

import { embedMany } from 'ai';
import { openai } from '@wix/ai';

const { embeddings } = await embedMany({
  model: openai.embeddingModel('text-embedding-3-small'),
  values: ['Hello', 'World', 'Foo', 'Bar'],
});

OpenAI Provider Methods

| Method | Description | |--------|-------------| | openai(modelId) | Responses API (default) | | openai.responses(modelId) | Responses API (explicit) | | openai.chat(modelId) | Chat Completions API | | openai.embeddingModel(modelId) | Text embeddings |

Runware Provider

Image Generation

import { generateImage } from 'ai';
import { runware } from '@wix/ai';

const result = await generateImage({
  model: runware.imageModel('runware:101@1'),
  prompt: 'A serene mountain landscape at sunset',
  n: 4, // Generate 4 images in one call
});

Using image() method

import { generateImage } from 'ai';
import { runware } from '@wix/ai';

const result = await generateImage({
  model: runware.image('FLUX.1'),
  prompt: 'A futuristic cityscape at night with neon lights',
});

Self-Hosted Apps

For apps running outside Wix, use createRunware with a WixClient:

import { generateImage } from 'ai';
import { createRunware } from '@wix/ai';
import { createClient, ApiKeyStrategy } from '@wix/sdk';

const wixClient = createClient({
  auth: ApiKeyStrategy({
    siteId: 'YOUR_SITE_ID',
    apiKey: 'YOUR_API_KEY',
  }),
});

const runware = createRunware({ client: wixClient });

const result = await generateImage({
  model: runware.imageModel('runware:101@1'),
  prompt: 'A beautiful sunset over mountains',
  n: 2,
});

Runware Provider Methods

| Method | Description | |--------|-------------| | runware.image(modelId) | Image generation (shortcut method) | | runware.imageModel(modelId) | Image generation model |

License

MIT