Seedream 5.0 Pro API: Layered Image Generation

Overview

Seedream 5.0 Pro is a ByteDance image generation model for high-quality text-to-image creation, reference-guided image generation, and layer decomposition from a single source image. It supports 1K and 2K generation tiers, flexible aspect ratios for generation and editing, optional explicit pixel sizes, and jpeg or png output.

Capability	Value
Model ID	`seedream-5-0-pro`
Modes	`text-to-image`, `image-to-image`, `layer-decomposition`
Prompt length	1-5000 characters; optional for `layer-decomposition`
Reference images	Required for `image-to-image`; required and exactly 1 public image URL for `layer-decomposition`
Images per task	1 output image for generation/editing; base image plus layers for `layer-decomposition`
Resolution tiers	`1K`, `2K` for generation/editing
Aspect ratios	`1:1`, `2:3`, `3:2`, `4:3`, `3:4`, `9:16`, `16:9`, `21:9` for generation/editing
Size input	Generation/editing: `1K`, `2K`, `1024x1024`, `2048X2048`, `2048*2048`; layer decomposition: `auto`, `1K`, `1.5K`, `2K`
Output formats	`jpeg`, `png`

Endpoint and authentication

Base URL:

https://api.apixo.ai/api/v1

Method	Endpoint	Purpose
`POST`	`/generateTask/seedream-5-0-pro`	Submit a generation task
`GET`	`/statusTask/seedream-5-0-pro?taskId={taskId}`	Poll task status and retrieve results

All requests require your APIXO API key:

Authorization: Bearer YOUR_API_KEY

Submit requests also require:

Content-Type: application/json

Copy-paste async quickstart

This minimal request submits a text-to-image task and returns a taskId.

curl -X POST "https://api.apixo.ai/api/v1/generateTask/seedream-5-0-pro" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "input": {
      "mode": "text-to-image",
      "prompt": "a cinematic product photo of wireless earbuds on black glass, crisp highlights",
      "resolution": "1K",
      "aspect_ratio": "1:1",
      "output_format": "jpeg"
    }
  }'

Successful response:

{
  "code": 200,
  "message": "success",
  "data": {
    "taskId": "task_12345678"
  }
}

Save the taskId; you need it to poll for the final result.

Poll for result

curl -X GET "https://api.apixo.ai/api/v1/statusTask/seedream-5-0-pro?taskId=task_12345678" \
  -H "Authorization: Bearer YOUR_API_KEY"

Success response:

{
  "code": 200,
  "message": "success",
  "data": {
    "taskId": "task_12345678",
    "state": "success",
    "resultJson": "{\"resultUrls\":[\"https://file.apixo.ai/xxx.jpeg\"]}",
    "createTime": 1767965610929,
    "completeTime": 1767965644132,
    "costTime": 33203
  }
}

Request body

Text-to-image

{
  "input": {
    "mode": "text-to-image",
    "prompt": "a clean product photo of wireless earbuds on soft stone",
    "resolution": "1K",
    "aspect_ratio": "1:1",
    "output_format": "jpeg"
  }
}

Image-to-image

{
  "input": {
    "mode": "image-to-image",
    "prompt": "keep the product shape, change the scene to a premium studio setup",
    "image_urls": [
      "https://example.com/reference-product.jpg"
    ],
    "resolution": "2K",
    "aspect_ratio": "4:3",
    "output_format": "png"
  }
}

Explicit size

{
  "input": {
    "mode": "text-to-image",
    "prompt": "a minimalist industrial poster with strong geometric composition",
    "size": "2048x2048",
    "output_format": "jpeg"
  }
}

Layer decomposition

{
  "input": {
    "mode": "layer-decomposition",
    "prompt": "decompose the product inside <bbox>10 20 600 900</bbox>",
    "image_urls": [
      "https://example.com/source-product.png"
    ],
    "size": "auto",
    "output_format": "png"
  }
}

For layer-decomposition, the model returns a base image plus extracted layer images. To guide the region to decompose, include a text tag such as <bbox>x1 y1 x2 y2</bbox> inside prompt. Do not send a separate bbox request field.

Parameters

string

default:"async"

Result delivery mode. Use async for polling with statusTask, or callback for webhook delivery.

string

Required when request_type is callback. Must be a public HTTPS URL that can receive the final task payload. See Webhooks.

object

required

Seedream 5.0 Pro input parameters.

Show properties

string

required

Generation mode. Supported values: text-to-image, image-to-image, layer-decomposition.

string

Text prompt describing the desired image. Required for text-to-image and image-to-image; optional for layer-decomposition. Supports 1-5000 characters after trimming whitespace. For layer-decomposition, include <bbox>x1 y1 x2 y2</bbox> in this text to guide the source-image region to decompose; there is no separate bbox request field.

string[]

Reference image URLs. Required for image-to-image. Accepts 1-10 public image URLs for image-to-image. Required for layer-decomposition and must contain exactly 1 public image URL. Each URL must be reachable by APIXO during task submission.

string

default:"1K"

Output resolution tier for text-to-image and image-to-image. Supported values: 1K, 2K. Not supported for layer-decomposition. When either resolution or aspect_ratio is provided, Seedream 5.0 Pro uses the resolution + aspect_ratio preset path; missing values default to 1K and 1:1.

string

default:"1:1"

Output aspect ratio for text-to-image and image-to-image. Supported values: 1:1, 2:3, 3:2, 4:3, 3:4, 9:16, 16:9, 21:9. Not supported for layer-decomposition. When resolution + aspect_ratio and size are both provided, resolution + aspect_ratio takes precedence.

string

Optional explicit size or preset token. For text-to-image and image-to-image, preset tokens support 1K and 2K; pixel strings accept x, X, or * separators, such as 1024x1024, 2048X2048, or 2048*2048. Custom pixel sizes must use 16-pixel granularity, keep aspect ratio between 1/16 and 16, and keep total pixels between 921,600 and 4,194,304 (1280x720 to 2048x2048 equivalent). For layer-decomposition, use one of auto, 1K, 1.5K, or 2K; default is auto.

string

default:"jpeg"

Output image format. Supported values: jpeg, png.

boolean

default:"false"

Enables content safety filtering for this request. Defaults to false. Must be a boolean when provided.

For layer-decomposition, do not send resolution, aspect_ratio, max_images, num_images, or n. The number of decomposition outputs is controlled by the model.

Size presets from `resolution + aspect_ratio`

When resolution or aspect_ratio is provided, Seedream 5.0 Pro maps the effective resolution + aspect_ratio pair to these output dimensions:

Aspect ratio	`1K` size	`2K` size
`1:1`	`1024x1024`	`2048x2048`
`4:3`	`1152x864`	`2304x1728`
`3:4`	`864x1152`	`1728x2304`
`16:9`	`1312x736`	`2848x1600`
`9:16`	`736x1312`	`1600x2848`
`3:2`	`1248x832`	`2496x1664`
`2:3`	`832x1248`	`1664x2496`
`21:9`	`1568x672`	`3024x1296`

Response format

Submit task response

POST /generateTask/seedream-5-0-pro returns a task ID when the task is accepted:

integer

API status code. 200 means the task was accepted.

Layer decomposition result

For mode=layer-decomposition, data.resultJson remains a compatibility JSON string containing only the resultUrls array. Read data.layerDecomposition for file-level decomposition details:

{
  "code": 200,
  "message": "success",
  "data": {
    "taskId": "task_12345678",
    "state": "success",
    "resultJson": "{\"resultUrls\":[\"https://file.apixo.ai/base.png\",\"https://file.apixo.ai/layer-1.png\"]}",
    "layerDecomposition": [
      {
        "role": "base",
        "url": "https://file.apixo.ai/base.png"
      },
      {
        "role": "layer",
        "url": "https://file.apixo.ai/layer-1.png",
        "size": "512x512",
        "output_format": "png",
        "z_index": 1,
        "bounding_box": {
          "absolute": [10, 20, 600, 900],
          "normalized": [5, 10, 293, 439]
        },
        "name": "foreground object",
        "description": "Extract the foreground object, no other elements included."
      }
    ],
    "createTime": 1783569600000,
    "completeTime": 1783569638000,
    "costTime": 38000
  }
}

string

Human-readable status message.

string

Task identifier used for later polling or webhook matching.

Latency and polling

Seedream 5.0 Pro tasks are asynchronous.

Output tier	Typical generation time	Recommended first poll	Poll interval
`1K`	20s-35s	20s after task creation	3s
`2K`	25s-50s	20s after task creation	3s

For production systems, prefer callback_url instead of tight polling loops.

Billing

Seedream 5.0 Pro uses output-image billing. text-to-image and image-to-image return one output image. layer-decomposition is quoted and charged as 4 expected output images because it returns a base image plus layer images.

Unit	APIXO price
Output up to and including 2.61M pixels	`$0.045 / output image`
Output above 2.61M pixels	`$0.09 / output image`
Additional input reference image after the first	`$0.003 / image`

Billing behavior:

output_charge = 1 * selected output pixel tier price
reference_charge = max(0, reference_image_count - 1) * 0.003
final charge = output_charge + reference_charge

For layer-decomposition, the request must contain exactly one input image, so the input reference add-on is $0. The default size=auto quote uses the 2K tier:

layer_decomposition_default_charge = 4 * 0.09 = 0.36

Explicit size=1K or size=1.5K uses the $0.045 output-image tier. Explicit size=2K uses the $0.09 output-image tier. Failed generations are refunded.

Practical notes

Image-count fields are intentionally omitted from this public contract for generation and editing; Seedream 5.0 Pro returns exactly one generated image for text-to-image and image-to-image.
3K and 4K are not supported by this endpoint.
2K + 21:9 maps to 3024x1296, which preserves the 21:9 ratio while staying within the 4,194,304 pixel limit.
If resolution or aspect_ratio is provided together with size, the effective resolution + aspect_ratio preset decides the output dimensions.
If you only provide size, APIXO uses that explicit pixel size or preset token.
For layer-decomposition, region control is text-based: write <bbox>x1 y1 x2 y2</bbox> in prompt; do not add a top-level or input-level bbox field.

​Overview

​Endpoint and authentication

​Copy-paste async quickstart

​Poll for result

​Request body

​Text-to-image

​Image-to-image

​Explicit size

​Layer decomposition

​Parameters

​Size presets from resolution + aspect_ratio

​Response format

​Submit task response

​Layer decomposition result

​Latency and polling

​Billing

​Practical notes

​Related links

Overview

Endpoint and authentication

Copy-paste async quickstart

Poll for result

Request body

Text-to-image

Image-to-image

Explicit size

Layer decomposition

Parameters

Size presets from `resolution + aspect_ratio`

Response format

Submit task response

Layer decomposition result

Latency and polling

Billing

Practical notes

Related links