Image Editor

Edit and transform existing images using text prompts with Google's Gemini 3 Pro model via fal.ai.

See references/fal-api.md for setup, Python patterns, and error handling.

API Endpoint

fal-ai/gemini-3-pro-image-preview/edit

Parameters

Required

prompt (string): The editing instruction describing what changes to make
image_urls (array of strings): URLs of the images to edit (use fal_client.upload_file() for local files)

Optional

Parameter	Type	Default	Options
`num_images`	integer	1	1-4
`aspect_ratio`	string	"auto"	"auto", "21:9", "16:9", "3:2", "4:3", "5:4", "1:1", "4:5", "3:4", "2:3", "9:16"
`output_format`	string	"png"	"jpeg", "png", "webp"
`resolution`	string	"1K"	"1K", "2K", "4K"
`sync_mode`	boolean	false	Returns data URI directly (no download needed)
`enable_web_search`	boolean	true	Grounds editing in current web data
`safety_tolerance`	string	"6"	Always set to "6" for maximum creative freedom
`seed`	integer	-	Fix seed for reproducible results
`limit_generations`	boolean	false	Restricts to 1 image per prompt round

Note: aspect_ratio defaults to "auto" for editing, which preserves the original image proportions.

CLI Script

python3 scripts/fal_generate.py \
    --endpoint edit \
    --prompt "Add snow to this mountain scene" \
    --image /path/to/mountain.jpg \
    --output edited.png

Tips

Be specific about what changes you want: "make the sky sunset orange" is better than "change the sky"
Use aspect_ratio: "auto" (the default) to preserve original image proportions
Multiple images can be provided for context, style transfer, or combining elements
Set enable_web_search: True for style-grounded edits ("make it look like a Monet painting")
Always include safety_tolerance: "6" for maximum creative flexibility
Upload local files with fal_client.upload_file() before passing to image_urls