GET

/v1/schemas/fal/image/fal-ai/fast-sdxl/image-to-image

kind inputactivity imagehash 7ef83f3c…extracted Nov 14, 2026version currentsource upstream spec ↗

Request body

19 properties · 2 required · 3 $defs
propertytypedescriptionconstraints
prompt*stringThe prompt to use for generating the image. Be as descriptive as possible for best results.
image_url*stringThe URL of the image to use as a starting point for the generation.
sync_modebooleanIf `True`, the media will be returned as a data URI and the output data won't be available in the request history.default: false
image_sizeImageSize | enumThe size of the generated image.default: "square_hd"
enable_safety_checkerbooleanIf set to true, the safety checker will be enabled.default: true
preserve_aspect_ratiobooleanIf set to true, the aspect ratio of the generated image will be preserved even if the image size is too large. However, if the image is not a multiple of 32 in width or height, it will be resized to the nearest multiple of 32. By default, this snapping to the …default: false
request_idstringAn id bound to a request, can be used with response to identify the request itself.default: ""
enable_prompt_expansionbooleanIf set to true, the prompt will be expanded with additional prompts.default: false
strengthnumberdetermines how much the generated image resembles the initial image0.05 ≤ n ≤ 1 · default: 0.95
safety_checker_versionenumThe version of the safety checker to use. v1 is the default CompVis safety checker. v2 uses a custom ViT model.v1 | v2 · default: "v1"
seedinteger The same seed and the same prompt given to the same version of Stable Diffusion will output the same image every time. nullable
negative_promptstringThe negative prompt to use.Use it to address details that you don't want in the image. This could be colors, objects, scenery and even the small details (e.g. moustache, blurry, low resolution).default: ""
formatenumThe format of the generated image.jpeg | png · default: "jpeg"
lorasLoraWeight[]The list of LoRA weights to use.default: []
crop_outputbooleanIf set to true, the output cropped to the proper aspect ratio after generating.default: false
embeddingsEmbedding[]The list of embeddings to use.default: []
num_inference_stepsintegerThe number of inference steps to perform.1 ≤ n ≤ 65 · default: 25
guidance_scalenumberThe CFG (Classifier Free Guidance) scale is a measure of how close you want the model to stick to your prompt when looking for a related image to show you.0 ≤ n ≤ 20 · default: 7.5
num_imagesintegerThe number of images to generate.1 ≤ n ≤ 8 · default: 1

Types

3 $defs

Embeddingobject

tokensstring[] The list of tokens to use for the embedding. When omitted, tokens are generated as <s0>, <s1>, ... to match the number of vectors in the embedding file. nullable
path*stringURL or the path to the embedding weights.

ImageSizeobject

widthintegerThe width of the generated image.0 < n ≤ 14142 · default: 512
heightintegerThe height of the generated image.0 < n ≤ 14142 · default: 512

LoraWeightobject

scalenumberThe scale of the LoRA weight. This is used to scale the LoRA weight before merging it with the base model.0 ≤ n ≤ 1 · default: 1
forcebooleanDeprecated. Accepted for backward compatibility and ignored.default: false
path*stringURL or path to a LoRA safetensors file. Or HF model name that resolves to safetensors weights.

Validate a payload

before you spend tokens
shell
$ curl -X POST https://modelschemas.com/v1/validate \
    -d '{"provider":"fal","endpointId":"fal-ai/fast-sdxl/image-to-image","payload":{…}}'
→ {"valid": false, "errors": [{"path": "#", "keyword": "required", …}]}