HomeDocsConsolePricing

text-to-image

Generates a brand-new image from a text description (no base image required). Reference images serve only as style/mood supplements, not as the primary content source. Covers general text-to-image generation, multi-image reference compositing, and commercial-grade high-quality generation with support for specified resolution (e.g., 4K wallpapers). Note: --image_list and --ratio are mutually exclusive; when --image_list is provided, --ratio is not supported.

meitu text-to-image

Usage Examples

# Scenario 1: Basic text-to-image (required parameters only)
meitu text-to-image \
  --prompt "An orange tabby cat sitting on a windowsill, minimalist illustration style" \
  --json

# Scenario 2: Reference image assisted + specified model
meitu text-to-image \
  --prompt "Futuristic city skyline" \
  --image_list ./mood_ref.jpg \
  --model praline_pro \
  --json

# Scenario 3: Specify resolution + download results
meitu text-to-image \
  --prompt "Alpine lake at sunrise, photorealistic style" \
  --size 4K \
  --ratio 16:9 \
  --json \
  --download-dir ./output

Parameter Reference

ParameterRequiredDescription
--image_listNoAliases: --image, --image_url
--promptYesType: string
--sizeNoType: string
--ratioNoType: string
--modelNoType: string
--download-dirNoType: string
--outputNoType: string[]
--jsonNoOutput results in JSON format for easy script or agent parsing