Seedream 4 API

Seedream 4 is a BytePlus image model designed for consistent visual storytelling. Use the Seedream 4 API to generate multi-image batches (up to 15 images per request, when enabled), blend style and identity with multi-reference inputs, and support high-resolution workflows for comics, ads, and product visuals. Availability and limits can vary by configuration—use your dashboard and API responses as the source of truth.

Prompt

A beautiful landscape with mountains and a lake, golden hour lighting, photorealistic style

Parameters

size

image_urls

Estimated Cost (per image)

1.8 Credits

Total = Base Price × Count (n)

Sample Result

Prompt

91 (suggested: 2,000)

Input Images

Upload reference images

Click to upload or drag and drop

Supported formats: JPG, JPEG, PNG, WEBP
Maximum file size: 10MB; Maximum files: 10

Size

For specific aspect ratios, describe them in your prompt

Count

Specifies the maximum number of images to generate (1–15).

History

Max 20 items

0 running · 0 completed

Your generation history will appear here

Pricing

Starting from

$0.025

1.8 Credits

per image

Guaranteed

99.9% Uptime

Seedream 4 API: The Engine for Consistent Visual Storytelling

Seedream 4 API is built for workflows that need consistency across a set of images—storyboards, character-led campaigns, and product visuals. Generate up to 15 images per request (when sequential generation is enabled), constrain identity/style with multi-reference inputs, and upscale outputs for high-resolution delivery. Exact results vary by prompt and configuration.

What can you build with Seedream 4 API?

Automated Storyboards

Create sequential narratives where characters and settings remain stable across shots using the `sequential_image_generation` parameter.

Start Using API

Consistent Product Visualization

Showcase products in varied environments without hallucinating details. Perfect for e-commerce integration via Seedream 4 API.

Start Using API

Adaptive Style Transfer

Apply specific artistic styles across disparate inputs. Ideal for maintaining brand identity in marketing campaigns.

Start Using API

Why developers choose Seedream 4 API

Designed for production pipelines requiring high coherence, speed, and massive context retention.

Unified Architecture

Perform Text-to-Image and Image-to-Image editing in a single model call, reducing latency and cost.

Deep Reference Fusion

Upload up to 10 reference images to heavily constrain style and identity, minimizing creative drift.

Production Speed

Seedream 4 is optimized for low-latency generation in many configurations. Actual latency varies by output size, batch settings, and routing—verify in your dashboard.

How to implement Seedream 4 API

A streamlined workflow for generating consistent image sequences.

Step 1 — Define References

Send a request to the Seedream 4 API with up to 10 image URLs to establish the 'Anchor' character or style.

Step 2 — Set Sequence Mode

Enable `sequential_image_generation='auto'` and define your narrative beats in the prompt to trigger batch creation.

Step 3 — Retrieve & Refine

Receive a JSON response with up to 15 consistent images. Use the editing endpoint to fix specific details if needed.

Open API Explorer

Seedream 4 API Capabilities

Technical specs for scalable image generation

Efficiency

15-Frame Batches

Generate entire scenes in one API call.

Control

Multi-Ref Fusion

Blend up to 10 inputs for precise control.

Consistency

Identity Lock

Maintains facial features across angles.

Quality

Native 4K

Output ultra-high resolution via EvoLink.

Editing

Smart Inpainting

Edit specific regions using natural language.

Speed

Low Latency

Designed for responsive generation in many configurations. Actual latency depends on size, batch count, and routing—verify in your dashboard.

Seedream 4 API vs Competitors

Choose the right model for your automated pipeline

Model	Category	Core capability	Output sizing	Billing	Best for
Seedream 4 (BytePlus)	Image generation + editing	Multi-image batches (up to 15), multi-reference fusion, consistent sequences	2K / 4K options (config dependent)	Per image (usage-based)	Storyboards, character consistency, and multi-image workflows that need stable visuals across shots
Nano Banana 2 (Gemini Flash Image)	Fast single-shot generation	Low-latency general generation	1024px base (config dependent)	Usage-based	Single-image generation where speed matters more than multi-frame consistency
Wan 2.5 (Text-to-Image)	Image generation	Strong prompt-following for creative generation (variant dependent)	1024–1536px (config dependent)	Usage-based	General creative generation when you don't need multi-frame sequence consistency

Seedream 4 API FAQs

Everything you need to know about the product and billing.

The API supports generating up to 15 coherent frames in a single batch request when using the sequential generation mode.

By passing multiple reference images (up to 10) of your character and using the unified generation endpoint, Seedream 4 locks identity features across the sequence.

Yes, its unified architecture allows for natural language editing (e.g., 'remove the hat') without needing separate masking tools.

Pricing is usage-based and billed per generated image. Effective cost can vary by size, batch settings, and routing; use your dashboard usage and billing records as the source of truth.

While Nano Banana (Gemini Flash Image) is optimized for sheer speed, Seedream 4 API prioritizes multi-frame consistency and complex instruction following for stories.

Usually no. The 'Reference Fusion' capability allows you to replicate styles instantly by uploading reference assets alongside your prompt.

API Reference

Select endpoint

Authentication

All APIs require Bearer Token authentication.

Header

Authorization: 
Bearer YOUR_API_KEY

Get API Key

POST

/v1/images/generations

Generate Image

Create an image generation task using text prompts or reference images. Supports text-to-image, image-to-image, and image editing modes.

Asynchronous processing mode, use the returned task ID to .

Generated image links are valid for 24 hours, please save them promptly.

Request Parameters

modelstringRequiredDefault: doubao-seedream-4.0

Image generation model name.

Exampledoubao-seedream-4.0

promptstringRequired

Prompt describing the image to be generated or how to edit the input image.

Notes

Limited to 2000 tokens

ExampleA cat playing in the grass

sizestringOptional

Size of generated images. Supports simplified format or pixel format.

Value	Description
1K	Standard quality
2K	High quality
4K	Ultra quality

Notes

Pixel format: 1280x720 to 4096x4096
Aspect ratio range: [1/16, 16]

Example2048x2048

nintegerOptionalDefault: 1

Number of images to generate.

Value	Description
1-15	Batch generation supported

Example1

image_urlsarrayOptional

Reference image URL list for image-to-image and image editing functions.

Notes

Max 10 images per request
Max size: 10MB per image
Formats: .jpeg, .jpg, .png
Aspect ratio range: [1/3, 3]
Total pixels: ≤6000×6000
URLs must be directly accessible by the server

Examplehttps://example.com/image1.png

prompt_priorityenum<string>OptionalDefault: standard

Prompt optimization strategy for setting the prompt optimization mode.

Value	Description
standard	Standard mode, higher quality, slower
fast	Fast mode, faster speed, normal quality

Examplestandard

callback_urlstringOptional

HTTPS callback address after task completion.

Notes

Triggered on completion, failure, or cancellation
Sent after billing confirmation
HTTPS only, no internal IPs
Max length: 2048 chars
Timeout: 10s, Max 3 retries

Examplehttps://your-domain.com/webhooks/image-task-completed

Request Example

{
  "model": "doubao-seedream-4.0",
  "prompt": "A cat playing on the grass",
  "size": "2K"
}

Response Example

{
  "created": 1757165031,
  "id": "task-unified-1757165031-uyujaw3d",
  "model": "doubao-seedream-4.0",
  "object": "image.generation.task",
  "progress": 0,
  "status": "pending",
  "task_info": {
    "can_cancel": true,
    "estimated_time": 45
  },
  "type": "image",
  "usage": {
    "billing_rule": "per_call",
    "credits_reserved": 1.8,
    "user_group": "default"
  }
}