Seedream 5.0 Pro

Seedream 5.0 Pro generates one image from a prompt, edits or combines up to 10 reference images, and splits one image into a base image plus up to 16 transparent layers. Output is 1K, 1.5K, 2K or an exact pixel size. Model details and Playground.

Connect to Wan APICopy heading link

Use https://api.wanapi.app as the API base URL. Create a key with the Wan model family in API key settings and send it as Authorization: Bearer YOUR_API_KEY. Keep keys on your server. The Playground uses your website session and charges the same account balance.

GET /v1/models lists the models available to your key. Seedream 5.0 Pro uses the OpenAI Images API: POST /v1/images/generations with JSON, or POST /v1/images/edits with JSON or multipart uploads. The request returns when the image is ready; there is no task to poll.

Generate an imageCopy heading link

Set WAN_API_KEY to your key on the server. Send the request once; generation usually finishes within a minute, so allow a client timeout of at least four minutes.

curl https://api.wanapi.app/v1/images/generations \
  -H "Authorization: Bearer $WAN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "bytedance/seedream-5.0-pro",
  "prompt": "A minimalist magazine cover: a red fox in fresh snow at dawn, the bold serif title \"QUIET WILD\" across the top",
  "size": "1K"
}'

The response follows the OpenAI Images format. Example values are illustrative:

{
  "created": 1792022400,
  "data": [
    {
      "url": "https://cdn.example.com/image-1.jpg",
      "size": "1024x1024",
      "output_format": "jpeg"
    }
  ],
  "usage": {
    "generated_images": 1
  }
}

Image URLs remain available for 30 days. Store the files yourself if you need them longer. Send "response_format": "b64_json" to also receive each image as Base64 in b64_json.

Seedream 5.0 Pro pricingCopy heading link

Each generated image is billed per image for its pixel size, at 30% below the official rate: up to 2.61 million pixels (1K and 1.5K) uses the lower rate, larger images the higher rate. The first reference image is free; each additional reference image adds the reference rate. A request rejected before an image is generated is not charged.

Layer decomposition bills the base image and every returned layer. When every returned image is up to 2.61 million pixels, each costs 30% below the official layer rate. When any returned image is larger, which is always the case at 2K, every image is charged at the official layer rate for its own size. Your balance must cover the maximum of 17 images; you pay only for the images returned.

Request fieldsCopy heading link

Send a JSON object. model is required; prompt is required unless layer_decomposition is true. Unknown fields are rejected.

stream, sequential_image_generation, tools, seed, guidance_scale and mask are not supported by this model and are rejected.

Edit and combine reference imagesCopy heading link

Pass one image with an edit instruction, or several images to combine:

{
  "model": "bytedance/seedream-5.0-pro",
  "prompt": "Place the perfume bottle from image 1 on the marble table from image 2, soft window light",
  "image": [
    "https://example.com/bottle.png",
    "https://example.com/table.jpg"
  ],
  "size": "2K",
  "output_format": "png"
}

POST /v1/images/edits accepts the same fields. With multipart, upload files in image or image[] fields and send optimize_prompt_options as a JSON string.

Layer decompositionCopy heading link

{
  "model": "bytedance/seedream-5.0-pro",
  "prompt": "Separate the title text, the parrot and the background",
  "image": "https://example.com/poster.png",
  "layer_decomposition": true,
  "size": "2K"
}

The response lists the base image first. Each layer adds its stacking order, position and description:

{
  "created": 1792022400,
  "data": [
    { "url": "https://cdn.example.com/image-1.jpg", "size": "2048x2048", "output_format": "jpeg", "z_index": 0 },
    {
      "url": "https://cdn.example.com/image-2.png",
      "size": "1273x265",
      "output_format": "png",
      "z_index": 1,
      "bounding_box": { "absolute": [383, 120, 1655, 384], "normalized": [187, 59, 808, 188] },
      "name": "Title text",
      "description": "Yellow serif title text"
    }
  ],
  "usage": { "input_images": 1, "generated_images": 2 }
}

If any layer fails, the whole request fails. If the prompt asks for more layers than the limit, some layer content may be dropped.

ErrorsCopy heading link

Errors use the OpenAI error envelope {"error": {"message", "type", "code"}}.

Never retry a request automatically after a timeout without checking your usage logs first.