Generating Images

Generating Images

Use AI models to generate images from text descriptions or reference images via OpenRouter API.

Quick start

Text-to-image generation

Generate an image from a text description:

python scripts/​generate.py \
  --prompt "A serene landscape with mountains and a lake at sunset" \
  --output "landscape.png"

Image-to-image generation

Generate a new image based on a reference image:

python scripts/​generate.py \
  --reference "girl.png" \
  --prompt "Same art style, but the girl is happily eating delicious braised pork" \
  --output "result.png"

Configuration

Initial setup

Set your OpenRouter API key as environment variable:

export OPENROUTER_API_KEY="${OPENROUTER_API_KEY}"

Aspect ratios

Specify custom aspect ratios with --aspect-ratio:

# Square image (1024x1024)
python scripts/​generate.py --prompt "..." --aspect-ratio "1:1"

# Landscape (1344x768)
python scripts/​generate.py --prompt "..." --aspect-ratio "16:9"

# Portrait (768x1344)
python scripts/​generate.py --prompt "..." --aspect-ratio "9:16"

Supported ratios: 1:1, 16:9, 9:16, 4:3, 3:4

Advanced features

Choose a different model

See MODELS.md for available models and their capabilities.

python scripts/​generate.py \
  --model "google/​gemini-2.5-flash-image" \
  --prompt "..." \
  --output "result.png"

More examples

See EXAMPLES.md for common use cases and scenarios.

Scripts reference

generate.py

Main image generation script.

Required arguments:

• --prompt: Text description of the image to generate
• --output: Output file path (e.g., result.png)

Optional arguments:

• --reference: Path to reference image for image-to-image generation
• --aspect-ratio: Image aspect ratio (default: 1:1)
• --model: Model to use (default: from config.json)
• --timeout: Request timeout in seconds (default: 60)

Examples:

Text-to-image:

python scripts/​generate.py \
  --prompt "Anime style girl drinking coffee" \
  --aspect-ratio "16:9" \
  --output "coffee.png"

Image-to-image:

python scripts/​generate.py \
  --reference "style_reference.png" \
  --prompt "Same style, different scene: girl eating ramen" \
  --output "ramen.png"

Error handling

The script handles common errors:

• Missing API key: Shows clear error message
• Network errors: Retries automatically
• Invalid image format: Validates before saving
• Missing reference file: Reports file not found

Check script output for detailed error messages.