Image Models

Available image generation models in the Decart provider

Model IDs

type DecartImageModelId = "lucy-pro-t2i";

The Decart provider currently supports one image generation model.

Available Models

lucy-pro-t2i

Text-to-image generation model powered by Lucy Pro.

Capabilities:

• Text-to-image generation
• Seed-based reproducibility
• Aspect ratio control (16:9 and 9:16)

Supported Settings:

• prompt - Text description of the image to generate
• seed - Random seed for reproducible generation
• aspectRatio - Only "16:9" (landscape) and "9:16" (portrait) are supported

Unsupported Settings:

• size - Will produce a warning if specified
• Other aspect ratios - Will produce a warning if specified

Usage

Basic Text-to-Image

import { createDecart } from '@decartai/ai-sdk-provider';
import { generateImage } from 'ai';

const decart = createDecart();

const { image } = await generateImage({
  model: decart.image('lucy-pro-t2i'),
  prompt: 'A serene mountain landscape at sunset'
});

With Seed for Reproducibility

import { createDecart } from '@decartai/ai-sdk-provider';
import { generateImage } from 'ai';

const decart = createDecart();

const { image } = await generateImage({
  model: decart.image('lucy-pro-t2i'),
  prompt: 'A futuristic cityscape',
  seed: 12345
});

With Aspect Ratio

import { createDecart } from '@decartai/ai-sdk-provider';
import { generateImage } from 'ai';

const decart = createDecart();

// Landscape orientation
const { image: landscape } = await generateImage({
  model: decart.image('lucy-pro-t2i'),
  prompt: 'Wide mountain vista',
  aspectRatio: '16:9'
});

// Portrait orientation
const { image: portrait } = await generateImage({
  model: decart.image('lucy-pro-t2i'),
  prompt: 'Tall skyscraper',
  aspectRatio: '9:16'
});

Model Specifications

specificationVersion · string · default: "v2"

All Decart image models implement the AI SDK v2 specification.

maxImagesPerCall · number · default: "1"

Each generation call produces exactly one image.

provider · string · default: "decart.image"

Provider identifier for the image models.

Response Format

Image models return binary image data with the following structure:

{
  images: [Uint8Array],
  warnings: Array<ImageModelV2CallWarning>,
  response: {
    modelId: string,
    timestamp: Date,
    headers: Record<string, string>
  }
}

Warnings

The model may return warnings for unsupported settings:

• unsupported-setting: size - When size parameter is provided
• unsupported-setting: aspectRatio - When an unsupported aspect ratio is provided (only 16:9 and 9:16 are supported)