Image Generation

The Vercel AI SDK supports generating images from text prompts using compatible models.

Generate images using experimental_generateImage().

import {
  experimental_generateImage as generateImage,
} from 'ai';

import {
  openai,
  IMAGE_MODEL,
} from './client';

async function makeImage() {
  const result = await generateImage({
    model: openai.image(IMAGE_MODEL),

    prompt:
      'A sunset over the mountains, digital art',

    n: 1,
    size: '1024x1024',
  });

  if (result.image?.base64) {
    console.log('Received Base64 image.');
  } else if (result.images?.[0]?.url) {
    console.log(result.images[0].url);
  }
}

Parameters

When using experimental_generateImage, the following parameters are supported:

ImageModel

required

The image model instance to use for generation.

string

required

The text description of the image to generate.

number

default:1

The number of images to generate.

string

default:"1024x1024"

The size of the generated image (e.g. "1024x1024").

​Image Generation

​Parameters

Image Generation

Parameters