Integration Guides
Errors
Pricing & Billing
Seedream Image Generation
Use CometAPI POST /v1/images/generations for Seedream text-to-image and image-to-image generation — supports sequential multi-image output and reference images.
{
"created": 123,
"data": [
{
"url": "<string>",
"b64_json": "<string>",
"size": "<string>"
}
],
"usage": {
"generated_images": 123,
"output_tokens": 123,
"total_tokens": 123
},
"model": "<string>"
}Overview
Generate images using ByteDance’s Seedream models through the standard OpenAI-compatible/v1/images/generations endpoint. Supports both text-to-image and image-to-image generation.
For the full model list, see the Models page. For detailed parameter documentation from the provider, see the VolcEngine documentation.
sequential_image_generation to auto and optionally configure sequential_image_generation_options.max_images to limit the number of output images. The total number of reference images plus generated images must be ≤ 15.Image-to-Image
Pass one or more reference images via theimage parameter (array of URLs or base64 data URIs) along with a text prompt. Supports up to 14 reference images.
Returned URL Gives AccessDenied?
url is a signed object-storage link that contains & and other special characters.If you copy the raw JSON and paste it directly into a browser or script, shell escaping or terminal truncation may strip the signature parameters, resulting in:
\{"Code":"AccessDenied","Message":"Access Denied","EC":"0003-00000015"\}
- curl: pipe through
| jq -r '.data[].url'to extract the URL cleanly. - Python: use
result.data[0].urlorimage["url"]— the string is already complete; pass it directly torequests.get()orwebbrowser.open(). - JavaScript: same idea —
result.data[0].url. - General: make sure
&and%in the URL are not truncated or double-escaped.
Authorizations
Bearer token authentication. Use your CometAPI key.
Body
Parameters for Seedream image generation (text-to-image and image-to-image).
Model to use. See the Models page for current Seedream model IDs.
Text prompt describing the image to generate. Supports both Chinese and English. Recommended max 300 Chinese characters or 1000 English characters.
Output format. url returns a temporary image link (valid for 24 hours); b64_json returns base64-encoded image data.
url, b64_json Output resolution. Presets 2K, 4K, or exact pixels like 2048x2048, 2304x1728, 2848x1600, 1600x2848, etc. Default: 2048x2048.
Reference image(s) for image-to-image generation. Each item is a public URL or base64 data URI (data:image/png;base64,...). Supports up to 14 reference images. Supported formats: jpeg, png, webp, bmp, tiff, gif.
When true, adds an 'AI Generated' watermark to the bottom-right corner of the output image.
Controls how closely the output follows the prompt. Higher values increase adherence but may reduce naturalness. Range: 1.0–10.0. Not supported by Seedream 5.0 (ignored).
Random seed for reproducible outputs. Range: -1 to 2147483647. Use -1 (default) for random generation. Legacy parameter — ignored by newer models.
Controls sequential (multi-image) generation mode.
auto: model decides whether to generate a sequence based on the promptdisabled: always generate a single image
auto, disabled Configuration for sequential (multi-image) generation. Only effective when sequential_image_generation is auto.
Show child attributes
Show child attributes
Output image file format. Options: png, jpeg.
png, jpeg Response
Successful Response
{
"created": 123,
"data": [
{
"url": "<string>",
"b64_json": "<string>",
"size": "<string>"
}
],
"usage": {
"generated_images": 123,
"output_tokens": 123,
"total_tokens": 123
},
"model": "<string>"
}