Image Generation

Generate images from text prompts, edit existing images, or inpaint with masks.

Ask AI about this page

Text to image

1egaki image "cinematic mountain village at sunrise" -o village.png

Specify a model with -m:

1egaki image "isometric floating city, detailed, soft colors" -m imagen-4.0-generate-001 -o city.png

Pass --input to provide a source image for editing:

1egaki image "add a red scarf and make it winter" --input portrait.jpg -o portrait-winter.png

Works with URLs too:

1egaki image "turn this into a manga panel" --input https://example.com/photo.jpg -o manga.png

Provide both --input and --mask to replace specific regions:

1234egaki image "replace the sky with dramatic storm clouds" \
  --input landscape.png \
  --mask mask.png \
  -o storm.png

The mask is a black-and-white image where white regions indicate areas to replace.

12egaki image "cyberpunk alley at night" --aspect-ratio 16:9
egaki image "polaroid-style travel photo" --aspect-ratio 4:5

Generate batch variations with -n:

12egaki image "minimal logo concepts for a cat cafe" -n 4 -o logo.png
# writes logo-0.png, logo-1.png, logo-2.png, logo-3.png

Models that support --seed produce the same image for the same seed:

1234egaki image "studio product shot of a ceramic mug" \
  -m imagen-4.0-generate-001 \
  --seed 42 \
  -o mug.png

Skip writing to disk and pipe directly to other tools:

1egaki image "flat icon of a fox" --stdout | magick - -resize 512x512 fox-icon.png

1egaki image "futuristic sneaker concept" -m gpt-image-1.5 --json -o sneaker.png

Prints a JSON object with the output path and metadata.

Model ID	Best for
`imagen-4.0-ultra-generate-001`	High-quality with seed + ratio
`gemini-3.1-flash-image-preview`	Fast, cheap text+image edits
`nano-banana-pro-preview`	Highest fidelity Google output
`gpt-image-1.5`	Strong editing and inpainting
`fal-ai/flux/schnell`	Very fast low-cost batches
`grok-imagine-image-quality`	xAI high quality with 2K output
`recraft-v4`	Design-focused generation

Run egaki models --type image to see all available image models with pricing.

Imagen (imagen-*): --seed, --aspect-ratio, --input, --mask, -n
Gemini: --input, --aspect-ratio, --image-size; usually no --seed
OpenAI: strong editing and inpainting; size controls are model-specific
BFL (flux-*): Kontext/Pro variants via AI Gateway subscription
Recraft (recraft-*): design-focused, v2/v3/v4 families
xAI (grok-imagine-*): --quality, --resolution, --output-format, --input