Seedream-5.0-Pro Image Generation
Generate high-quality images using Seedream 5.0 Pro model with 1K/2K resolution
Note for users in mainland China: Please use https://toapis.cn as the API endpoint (Base URL). Replace https://toapis.com with https://toapis.cn in the examples in this document.
- Seedream 5.0 Pro high-quality image generation model from ByteDance Seed team
- Use the model parameter to select
doubao-seedream-5-0-pro - Parameters are largely the same as
Seedream-5.0, but Pro supports 1K and 2K resolutions - Supports text-to-image, image-to-image, and multi-reference generation
- Billing rule: the first reference image is free; billing starts from the second reference image
- Async task management, query results by task ID
Important Change: For better performance and cost control, we no longer support passing base64 image data directly in image_urls. Please use the Upload Image API first to upload images and get URLs, then call this endpoint.
Authorizations
AuthorizationstringheaderrequiredAll endpoints require Bearer Token authentication
Get your API Key from the API Key Management Page
Add to request header:
Authorization: Bearer YOUR_API_KEYBody
modelstringbodyrequireddefault: doubao-seedream-5-0-proImage generation model name
Example: "doubao-seedream-5-0-pro"
promptstringbodyrequiredText description for image generation
sizestringbodydefault: 1:1Image aspect ratio
When layer_decomposition is enabled, pass the 1K or 2K resolution preset directly in this field.
Supported aspect ratios:
1:1- Square (default)4:3- Landscape 4:33:4- Portrait 3:416:9- Landscape widescreen9:16- Portrait vertical3:2- Landscape 3:22:3- Portrait 2:321:9- Ultra-wide9:21- Ultra-tall
nintegerbodydefault: 1Number of images to generate
Range: 1-10 (minimum 1, maximum 10)
Default: 1
Note:
- The total of reference images + generated images must be ≤ 15
- Must enter a plain number (e.g.,
1), do not use quotes or it will cause an error
Charges will be pre-deducted based on generated image count and reference image count, with the first reference image free
layer_decompositionbooleanbodydefault: falseEnables the layer decomposition capability available only in Seedream 5.0 Pro.
false: returns a regular flattened image (default)true: returns a base image plus independent transparent PNG layers
When enabled, each result.data[] item includes z_index; independent layers also include bounding_box, name, and description.
image_urlsstring[]bodyReference image URL list for image-to-image or image editing
⚠️ URL format only (base64 no longer supported)
- Publicly accessible image URL (http:// or https://)
- You can use the Upload Image API to upload local images and get URLs
Limitations:
- Image count: 1-10
- Image formats: jpeg, png
- Aspect ratio (width/height) range: [1/3, 3]
- Width and height (px) > 14
- Size: not exceeding 10MB
- Total pixels: not exceeding 6000×6000 px
Billing notes:
- The first reference image is free
- Billing applies from the second reference image onward
metadataobjectbodyChannel-specific parameters for advanced Seedream 5.0 Pro model configuration
metadata properties
metadata.resolutionstringbodydefault: 2KImage resolution
Supported resolutions:
1K- Basic resolution2K- Standard resolution (default)
Note: Seedream 5.0 Pro does not support 3K or 4K resolution; non-Pro Seedream 5.0 supports 2K and 3K
metadata.sequential_image_generationstringbodydefault: disabledSequential image generation mode (Doubao-specific feature)
Controls whether to generate multiple images:
disabled: Disable sequential mode, generates only 1 image even with multiple reference images (default)auto: Enable sequential mode, can generate multiple images
Usage Notes:
- ✅ Set
n: 3or usesequential_image_generation: "auto"+max_images: 3 - ✅ Supports text-to-image sets: up to 10 images from text prompt only
- ✅ Supports image-to-image sets: provide
image_urlsto generate multiple variations - ⚠️ When
n > 1, it will automatically be set toauto
Note:
- Total reference images + generated images must be ≤ 15
metadata.sequential_image_generation_optionsobjectbodySequential image generation options
Available when sequential_image_generation is set to auto
Properties:
max_images(integer): Specify the number of images to generate, Range: 1-10
Example:
"sequential_image_generation_options": { "max_images": 3 }metadata.watermarkbooleanbodydefault: falseWhether to add a watermark to the generated image
true: Add watermarkfalse: No watermark (default)
Response
idstringUnique task identifier for status queries
objectstringObject type, always generation.task
modelstringModel name used
statusstringTask status
queued- Queued for processingin_progress- Processingcompleted- Successfully completedfailed- Failed
progressintegerTask progress percentage (0-100)
created_atintegerTask creation timestamp (Unix timestamp)
metadataobjectTask metadata
completed_atintegerTask completion timestamp (Unix timestamp). Returned only after the task completes.
expires_atintegerResult file expiration timestamp (Unix timestamp). Download or persist the files before they expire.
resultobjectGenerated result returned when status is completed.
result.typestringResult type. Always image for image tasks.
result.dataarrayImage result list. With layer decomposition enabled, the first item is normally the base image and the remaining items are independent transparent PNG layers.
result.data[].urlstringPublic URL of the base image or independent layer.
result.data[].z_indexintegerLayer stacking order. With layer decomposition enabled, the base image is normally 0 and independent layers increase from there.
result.data[].sizestringPixel dimensions of the base image or independent layer, for example 2048x2048.
result.data[].output_formatstringFile format of this result item. The base image can use the requested format; transparent independent layers use png.
result.data[].bounding_boxobjectLayer position on the base image, with absolute and normalized coordinates.
result.data[].bounding_box.absolutearrayAbsolute pixel coordinates on the base image in [x1, y1, x2, y2] order, representing the top-left and bottom-right corners.
result.data[].bounding_box.normalizedarrayNormalized coordinates in [x1, y1, x2, y2] order, with values ranging from 0 to 1000.
result.data[].namestringLayer name.
result.data[].descriptionstringDescription of the layer contents.
requestExample
curl --request POST \
--url https://toapis.com/v1/images/generations \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '{
"model": "doubao-seedream-5-0-pro",
"prompt": "A cute panda playing in a bamboo forest",
"size": "1:1",
"n": 1,
"metadata": {
"resolution": "2K"
}
}'curl --request POST \
--url https://toapis.com/v1/images/generations \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '{
"model": "doubao-seedream-5-0-pro",
"prompt": "A cute panda playing in a bamboo forest",
"size": "1:1",
"n": 1,
"image_urls": [
"https://example.com/panda.jpg"
],
"metadata": {
"resolution": "1K"
}
}'curl --request POST \
--url https://toapis.com/v1/images/generations \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '{
"model": "doubao-seedream-5-0-pro",
"prompt": "Precisely separate the image into a base image and independent transparent PNG layers.",
"image_urls": [
"https://example.com/office-desk.png"
],
"size": "2K",
"n": 1,
"layer_decomposition": true
}'When layer_decomposition is enabled, set size directly to the 1K or 2K resolution preset. Do not use an explicit WIDTHxHEIGHT pixel size. After submission, query GET /v1/images/generations/{task_id} for the complete layer result.
curl --request POST \
--url https://toapis.com/v1/images/generations \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '{
"model": "doubao-seedream-5-0-pro",
"prompt": "A cute panda playing in a bamboo forest",
"size": "1:1",
"n": 4,
"metadata": {
"resolution": "2K",
"sequential_image_generation": "auto",
"sequential_image_generation_options": { "max_images": 4 },
"watermark": false
}
}'import requests
response = requests.post(
"https://toapis.com/v1/images/generations",
headers={
"Authorization": "Bearer your-ToAPIs-key",
"Content-Type": "application/json"
},
json={
"model": "doubao-seedream-5-0-pro",
"prompt": "A cute panda playing in a bamboo forest",
"size": "1:1",
"n": 1,
"metadata": {
"resolution": "2K"
}
}
)
task = response.json()
print(f"Task ID: {task['id']}")
print(f"Status: {task['status']}")import requests
response = requests.post(
"https://toapis.com/v1/images/generations",
headers={
"Authorization": "Bearer your-ToAPIs-key",
"Content-Type": "application/json"
},
json={
"model": "doubao-seedream-5-0-pro",
"prompt": "A cute panda playing in a bamboo forest",
"size": "1:1",
"n": 1,
"image_urls": ["https://example.com/panda.jpg"],
"metadata": {
"resolution": "1K"
}
}
)
task = response.json()
print(f"Task ID: {task['id']}")
print(f"Status: {task['status']}")const response = await fetch('https://toapis.com/v1/images/generations', {
method: 'POST',
headers: {
'Authorization': 'Bearer your-ToAPIs-key',
'Content-Type': 'application/json'
},
body: JSON.stringify({
model: 'doubao-seedream-5-0-pro',
prompt: 'A cute panda playing in a bamboo forest',
size: '1:1',
n: 1,
metadata: {
resolution: '2K'
}
})
});
const task = await response.json();
console.log(`Task ID: ${task.id}`);
console.log(`Status: ${task.status}`);const response = await fetch('https://toapis.com/v1/images/generations', {
method: 'POST',
headers: {
'Authorization': 'Bearer your-ToAPIs-key',
'Content-Type': 'application/json'
},
body: JSON.stringify({
model: 'doubao-seedream-5-0-pro',
prompt: 'A cute panda playing in a bamboo forest',
size: '1:1',
n: 1,
image_urls: ['https://example.com/panda.jpg'],
metadata: {
resolution: '1K'
}
})
});
const task = await response.json();
console.log(`Task ID: ${task.id}`);
console.log(`Status: ${task.status}`);responseExample
{
"id": "task_img_abc123def456",
"object": "generation.task",
"model": "doubao-seedream-5-0-pro",
"status": "queued",
"progress": 0,
"created_at": 1703884800,
"metadata": {}
}{
"completed_at": 1787128810,
"created_at": 1787128650,
"expires_at": 1789720810,
"id": "tsk_img_example_layer_decomposition",
"model": "doubao-seedream-5-0-pro",
"object": "generation.task",
"progress": 100,
"result": {
"type": "image",
"data": [
{
"url": "https://files.toapis.com/images/example/base.jpg",
"size": "1600x2848",
"output_format": "jpeg",
"z_index": 0
},
{
"url": "https://files.toapis.com/images/example/layer-1.png",
"size": "3040x955",
"output_format": "png",
"z_index": 1,
"bounding_box": {
"absolute": [0, 0, 1600, 503],
"normalized": [0, 0, 999, 176]
},
"name": "Background office equipment",
"description": "The monitor edge and file organizer extracted from the background"
},
{
"url": "https://files.toapis.com/images/example/layer-2.png",
"size": "1982x1162",
"output_format": "png",
"z_index": 2,
"bounding_box": {
"absolute": [800, 589, 1599, 1058],
"normalized": [500, 207, 999, 371]
},
"name": "Stacked white papers",
"description": "Several stacked white office papers extracted from the desk"
}
]
},
"status": "completed"
}