Image-to-image with a mask
Selectively modify portions of an image using a mask. The mask must be the same shape and size as the init image. This endpoint also supports image parameters with alpha channels. See below for more details.
NOTE: Only Version 1 engines will work with this endpoint.
Authorizations
Your Stability API key, sent as a bearer token in the Authorization header.
Headers
The format of the response. Leave blank for JSON, or set to 'image/png' for a PNG image.
application/json, image/png Allows for requests to be scoped to an organization other than the user's default. If not provided, the user's default organization will be used.
Used to identify the source of requests, such as the client application or sub-organization. Optional, but recommended for organizational clarity.
Used to identify the version of the application or service making the requests. Optional, but recommended for organizational clarity.
Path Parameters
Body
- Option 1
- Option 2
Represents the optional parameters that can be passed to any generation request.
An array of text prompts to use for generation.
Due to how arrays are represented in multipart/form-data requests, prompts must adhere to the format text_prompts[index][text|weight],
where index is some integer used to tie the text and weight together. While index does not have to be sequential, duplicate entries
will override previous entries, so it is recommended to use sequential indices.
Given a text prompt with the text A lighthouse on a cliff and a weight of 0.5, it would be represented as:
To add another prompt to that request simply provide the values under a new index:
1Image used to initialize the diffusion process, in lieu of random noise.
For any given pixel, the mask determines the strength of generation on a linear scale. This parameter determines where to source the mask from:
MASK_IMAGE_WHITEwill use the white pixels of the mask_image as the mask, where white pixels are completely replaced and black pixels are unchangedMASK_IMAGE_BLACKwill use the black pixels of the mask_image as the mask, where black pixels are completely replaced and white pixels are unchangedINIT_IMAGE_ALPHAwill use the alpha channel of the init_image as the mask, where fully transparent pixels are completely replaced and fully opaque pixels are unchanged
Optional grayscale mask that allows for influence over which pixels are eligible for diffusion and at what strength. Must be the same dimensions as the init_image. Use the mask_source option to specify whether the white or black pixels should be inpainted.
How strictly the diffusion process adheres to the prompt text (higher values keep your image closer to your prompt)
0 <= x <= 357
FAST_BLUE, FAST_GREEN, NONE, SIMPLE, SLOW, SLOWER, SLOWEST "FAST_BLUE"
Which sampler to use for the diffusion process. If this value is omitted we'll automatically select an appropriate sampler for you.
DDIM, DDPM, K_DPMPP_2M, K_DPMPP_2S_ANCESTRAL, K_DPM_2, K_DPM_2_ANCESTRAL, K_EULER, K_EULER_ANCESTRAL, K_HEUN, K_LMS "K_DPM_2_ANCESTRAL"
Number of images to generate
1 <= x <= 101
Random noise seed (omit this option or use 0 for a random seed)
0 <= x <= 42949672950
Number of diffusion steps to run.
10 <= x <= 5050
Pass in a style preset to guide the image model towards a particular style. This list of style presets is subject to change.
enhance, anime, photographic, digital-art, comic-book, fantasy-art, line-art, analog-film, neon-punk, isometric, low-poly, origami, modeling-compound, cinematic, 3d-model, pixel-art, tile-texture Extra parameters passed to the engine. These parameters are used for in-development or experimental features and may change without warning, so please use with caution.
Response
Generation successful.
An array of results from the generation request, where each image is a base64 encoded PNG.