Skip to main content
POST
/
v1
/
images
/
edits
curl https://api.cometapi.com/v1/images/edits \
  -H "Authorization: Bearer $COMETAPI_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "doubao-seededit-3-0-i2i-250628",
    "image": "https://your-image-host/source.jpg",
    "prompt": "Add warm sunset light",
    "response_format": "url"
  }'
{
  "model": "<string>",
  "created": 123,
  "data": [
    {
      "url": "<string>"
    }
  ],
  "usage": {
    "generated_images": 123,
    "output_tokens": 123,
    "total_tokens": 123
  }
}

​
Overview

Edit images using ByteDance’s SeedEdit model through the standard OpenAI-compatible /v1/images/edits endpoint. Apply prompt-guided edits, masks, and variations. For detailed parameter documentation, see the VolcEngine SeedEdit documentation.

​
When to use it

Use this endpoint when you already have an image and want to change part of it with a prompt. Typical cases include product image cleanup, background edits, local style changes, and prompt-guided variations.

​
First request

Start with one source image and a direct edit prompt. Add a mask only after the basic edit works, because masks introduce another failure point in file upload, transparency, and alignment.

​
Read the response

The API playground shows the exact response fields for generated images. Save the returned image URL or base64 payload according to your storage policy before you show it to users.

Authorizations

​
Authorization
string
header
required

Bearer token authentication. Use your CometAPI key.

Body

​
image
file
required

Incoming images need to fulfill the following conditions: Image format: jpeg, png. Aspect ratio (width/height): in the range (1/3, 3). Aspect length (px) > 14. Size: no more than 10MB.

​
prompt
string
required

A text description of the desired image(s).

  • Max length: 1000 chars for dall-e-3, 32000 chars for gpt-image-1.
Example:

"Put on glasses"

​
model
string

Model to use, such as doubao-seededit-3-0-i2i-250628. Choose a current image editing model from the Models page.

Example:

"doubao-seededit-3-0-i2i-250628"

​
response_format
enum<string>

Response format. Options: "url" or "b64_json". Only "dall-e-3" supports "url" (valid for 60 minutes).

Available options:
url,
b64_json
​
size
string

Output image size. Use adaptive to match the input image dimensions.

Example:

"adaptive"

​
seed
integer
default:-1

Random seed for reproducible outputs. Range: -1 to 2147483647. Use -1 (default) for random generation.

​
guidance_scale
number

Controls how strongly the prompt influences the result vs. the input image. Range: 1.0–10.0. Higher values follow the prompt more closely.

​
watermark
boolean
default:true

When true, adds an 'AI Generated' watermark to the bottom-right corner of the output image.

Response

200 - application/json

success

​
model
string
required
​
created
integer
required
​
data
object[]
required
​
usage
object
required