Edit an image

Create an edited result from one or more input images and a text prompt. You can also provide a mask to identify the area to change.

POST/v1/images/edits

Create an edited image based on one or more input images and text prompts. When providing a mask, the transparent area in the mask indicates where editing is needed; When editing multiple images, the mask is applied to the first input image.

The input file type, size, quantity, mask rules, as well as quality and size options depend on the selected model. Please use the model ID visible and supported by the current API Key for image editing

Authorization

bearerAuth

AuthorizationBearer <token>

Use Bearer your-api-key in the Authorization Header.

In: header

Request Body

multipart/form-data

Use Multipart to submit images, prompts, and optional masks. The client should automatically generate the boundary of the Content-Type from the request library.

TypeScript Definitions

Use the request body type in TypeScript.

image*|array<>

One or more images to edit DALL· E 2 accepts square PNG files smaller than 4 MB; If no mask is provided, the image must include transparent areas. GPT-Image-1 can accept multiple PNGs, WEBPs, or JPGs, each less than 25 MB. The actual supported file formats, sizes, and quantities of depend on the selected model

prompt*string

Expected editing results. DALL· E 2 accepts up to 1,000 characters, GPT-Image-1 accepts up to 32,000 characters; Other models are limited by their actual capabilities.

Length1 <= length

mask?file

Optional mask image. The fully transparent area indicates the location to be edited. The mask should be a PNG less than 4 MB and the same size as the input image; For multi-image editing, the mask is applied to the first image.

Formatbinary

model*string

The image editing model ID to use. Please select the model that is visible to the current API Key and supports image editing.

Length1 <= length

n?integer|null

Number of images to return, ranging from 1 to 10; the actual supported number depends on the selected model.

Range1 <= value <= 10

Default1

quality?string|null

Image quality. GPT-Image-1 supports high, medium, low, DALL· E 2 uses standard, with the default value being auto.

Default"auto"

Value in

"auto"
"high"
"medium"
"low"
"standard"
null

response_format?string|null

Image response format. url returns an image URL that remains valid for 60 minutes after generation; b64_json returns Base64-encoded image data.

Default"url"

Value in

"url"
"b64_json"
null

size?string|null

Image size. GPT-Image-1 supports 1024x1024, 1536x1024, 1024x1536, and auto, while DALL· E 2 supports 256x256, 512x512, and 1024x1024.

Default"1024x1024"

Value in

"256x256"
"512x512"
"1024x1024"
"1536x1024"
"1024x1536"
"auto"
null

user?string

Unique identifier for the end user, used to help monitor and detect abuse.

Length1 <= length

Response Body

`application/json`

curl -X POST "https://example.com/v1/images/edits" \  -F model="MODEL_ID" \  -F image="otter.png" \  -F mask="mask.png" \  -F prompt="A cute little sea otter wearing a beret" \  -F n="1" \  -F size="1024x1024"

{
  "created": 1589478378,
  "data": [
    {
      "url": "https://example.com/edited-image-1.png"
    },
    {
      "url": "https://example.com/edited-image-2.png"
    }
  ]
}

Authorization

Request Body

Response Body

200application/json

400application/json

401application/json

403application/json

`application/json`

`application/json`

`application/json`

`application/json`