falImage to imageComing soon

bria/fibo-edit-1.5/expand

Name: bria/fibo-edit-1.5/expand
Brand: fal

bria/fibo-edit-1.5/expand is an image to image tool from fal that runs in the background and returns images. Coming soon: listed with its price, not yet open to runs.

This tool is listed but not yet available to run.

Priced by usage

Pricing

$0.052 per image. The final charge comes from usage.

Limits

Runs: In the background
Returns: image
Longest queue wait: 3600s
Run deadline: 1800s
Largest inline input: 1024 KB
Largest input file: 500 MB
Largest output: 4768 MB
Outputs per run: 16

Input

Field	Kind	Default	Description
`image_url` required	URL or file	—	The image to expand (file or URL).
`canvas_size`	list	—	The desired size of the final image after the expansion, as [width, height]. Area must be under 25 megapixels. Pairs with 'original_image_location'.
`aspect_ratio`	one of 1:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9	—	The desired aspect ratio of the final image. Used over 'original_image_size' and 'original_image_location' if provided: the original is centred and the canvas grows around it.
`original_image_size`	list	—	The desired size of the original inside the canvas, as [width, height]; it is resized to this before being placed. Ignored if 'aspect_ratio' is provided.
`original_image_location`	list	—	Top-left corner of the original within the canvas, as [x, y]. May sit outside the canvas, in which case the original is cropped. Ignored if 'aspect_ratio' is provided.
`prompt`	text	—	What should appear in the expanded area. Omit it and the model is asked only to extend the image, which is usually what you want for a plain canvas change.
`seed`	number (whole) 0 to 2147483647	—	Reuse the seed from a previous response to recreate that result. Omit it for a random one, which is reported in the response either way.
`negative_prompt`	text	—	Not supported by this model, and refused rather than ignored -- see 'prompt'. Present only so a request written against the SDXL expansion endpoint fails loudly here instead of silently losing the field.
`retain_original_quality`	true or false	true	Keep the original's own pixels at full resolution. Generation happens near one megapixel, so a larger canvas is scaled up afterwards; leave this on and the region the mask protected is not resampled at all. Turn it off for a uniformly soft result, which can hide the seam on a textured background.
`preserve_alpha`	true or false	true	Keep the input's transparency. The model works in RGB, so the alpha channel is not generated -- the original's own is extended outwards by replicating its edges, which is what expanding a cutout means. Turn it off to get a flattened RGB result.
`return_model_output`	true or false	false	Also return the model's own frame, before the original was pasted back and the seam blended. Diagnostic only -- it does not change the returned image. Use it to tell a bad generation apart from a bad composite: a boundary already broken in that frame came from the model, not from post-processing.

Same request in code

curl https://api.spiral.ai/v1/runs \
  -H "Authorization: Bearer $SPIRAL_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $(uuidgen)" \
  -d '{
    "provider": "fal",
    "model": "bria/fibo-edit-1.5/expand",
    "input": {
      "image_url": "https://storage.googleapis.com/falserverless/model_tests/orange.png"
    }
  }'