01
Seedream 4 Text to Image
Use `seedream-4` for standard generation. Reference images are optional, so you can go from pure prompts to reference-guided generation without changing endpoints.
seedream-4
Text to Image
Image to Image
Model:
Seedream 4 API supports text-to-image and image editing for creative assets, product visuals, marketing images, and automated design workflows.
Input
836/5000
Added 0/14
Drop files here, paste files, or add a URL. seedream-4 can use reference images optionally. seedream-4-edit requires them. Current limit: 14 image(s).
Output
Example Output
This is sample data. Generate a new result to see real output.
Task ID:seedream-4-previewCreated:2026-06-09T00:00:00
Status:finishedProgress:-
Examples
Pricing details
Transparent pricing with no hidden fees. Pay as you go.
seedream-4Text to Image
PoYo price
$0.025/gen
5 credits
Official price
-
You save
-
seedream-4-editImage Edit
PoYo price
$0.025/gen
5 credits
Official price
-
You save
-
| Model | Spec | PoYo price | Official price | You save |
|---|---|---|---|---|
| seedream-4 | Text to Image | $0.025/gen 5 credits /gen | - | - |
| seedream-4-edit | Image Edit | $0.025/gen 5 credits /gen | - | - |
* Actual fees are based on the final output.
Introduction
Complete guide to using Seedream 4 API - Native 1K to 4K Image Creation on PoYo
Seedream 4 API - Native 1K to 4K Image Creation on PoYo
Access Seedream 4 API instantly on PoYo with one async workflow for text-to-image and image editing. Choose `1K`, `2K`, or `4K`, set `n` from `1` to `15`, and add reference images when you need stronger composition or identity control.
Available Seedream 4 API Models on PoYo
Get API Key
02
Seedream 4 Image Edit
Use `seedream-4-edit` when you need direct editing. `image_urls` is required, and the same endpoint returns an async task you can poll through the standard status API.
Key Features of Seedream 4 API
Built from ByteDance Seedream 4.0 with unified generation, editing, and multi-image reasoning.
01
One Workflow for Generation and Editing
Seedream 4 combines text-to-image generation and image editing in one API family. On PoYo you submit both modes through the same async endpoint, which keeps integration simple when your app needs prompt-only creation and reference-based editing in the same product.
- Use one submit endpoint
- Switch between generation and editing by model name
- Keep the same polling workflow for results
02
Native 1K, 2K, and 4K Output
The PoYo API exposes `resolution` as `1K`, `2K`, or `4K`, plus standard aspect-ratio `size` values. Seedream 4 also supports `n` from `1` to `15`, which makes it useful for ideation, batch asset generation, and prompt A/B testing from one request.
- Pick 1K, 2K, or 4K
- Batch up to 15 outputs per call
- Use common image aspect ratios directly
03
Reference-Guided Reasoning and Composition
ByteDance's Seedream 4.0 technical report describes a unified system for text-to-image, editing, and multi-image composition. In practice, that matters when you need the model to preserve visual identity, combine several references, or follow composition-heavy prompts instead of producing generic single-shot outputs.
- Use references for subject consistency
- Support multi-image composition tasks
- Handle prompt-sensitive layouts more reliably
Who Can Benefit from Seedream 4 API?
01
Product and Catalog Teams
Generate multiple product variations at once, move from 1K drafts to 4K finals, and add references when packaging, colors, or layouts need to stay closer to the source brief.
02
Design and Poster Workflows
Use Seedream 4 for poster comps, key visuals, and design exploration where batch output matters more than a single final image. The `n` parameter is useful for broad prompt exploration.
03
Editing Pipelines
Plug `seedream-4-edit` into apps that need reference-based image changes without building a separate editing backend. Upload source images, submit the edit task, then poll for completion.
04
Educational and Knowledge Visuals
Seedream 4's multi-image reasoning is useful for diagrams, explainers, and knowledge-heavy assets where the model needs to combine reference context instead of inventing everything from scratch.
Seedream 4 vs Seedream 4.5 vs Seedream 5.0 Lite - Model Comparison
Comparison based on PoYo API docs plus official ByteDance Seed technical materials. Use this table to choose between batch-friendly Seedream 4, more refined Seedream 4.5, and web-aware Seedream 5.0 Lite.
| Feature | Seedream 4 | Seedream 4.5 | Seedream 5.0 Lite |
|---|---|---|---|
| Resolution | 1K, 2K, 4K | 2K, 4K | 2K, 3K |
| Batch Output | Yes, n = 1 to 15 | Single output on current PoYo page | Single output on current PoYo page |
| Reference Images | Optional in text mode, required in edit mode | Supported | Supported |
| Web Search | No | No | Yes |
| Best For | Batch ideation and reference-guided creation | Higher consistency and denser typography | Current-topic visuals and smart editing |
| Price on PoYo | 5 credits | 5 credits | 5 credits |
Seedream 4.5 official materials state that 4.5 improves prompt adherence, alignment, and aesthetics over Seedream 4.0.
How to Use Seedream 4 API on PoYo
Step 1: Create your API key
Sign up on PoYo and generate a key for the standard async image workflow.
Create API Key →
Step 2: Set the right request parameters
Choose `seedream-4` or `seedream-4-edit`, set `size`, `resolution`, and `n`, then add `image_urls` if you want reference-guided creation or editing.
Add Credits →
Step 3: Submit and poll the task
Send the request to the unified submit endpoint, then query the task status API until the result is `finished`. Edit mode requires `image_urls`, and the total of `image_urls + n` must stay within `15`.
View Documentation →
Seedream 4 API Pricing: Flexible 4K Generation
Seedream 4 API: $0.025 per image (5 credits) on PoYo.
Use the same pricing baseline for text-to-image and image editing, then scale output volume with `n`. PoYo pre-deducts credits by `n`, and the docs note that unused output credits are refunded automatically if fewer images are returned.
No subscription fees. Pay only for what you use.
Frequently Asked Questions about Seedream 4 API
What parameters does Seedream 4 API support on PoYo?
Seedream 4 supports `size`, `resolution`, `n`, and optional `image_urls` in text mode. `seedream-4-edit` requires `image_urls`. `resolution` supports `1K`, `2K`, and `4K`; `n` supports `1` to `15`; and the total of `image_urls + n` must not exceed `15`.
Can I use reference images with seedream-4 text-to-image?
Yes. The PoYo docs explicitly state that `seedream-4` supports text-to-image generation with optional reference images, so you can push the model closer to a target subject, style, or layout without switching to edit mode.
What is the difference between seedream-4 and seedream-4-edit?
`seedream-4` is the standard text-to-image model and can optionally take references. `seedream-4-edit` is the edit-focused model and requires `image_urls`, which makes it the better fit when a source image is the center of the workflow.
How does Seedream 4 compare with Seedream 4.5?
Official ByteDance Seed materials say Seedream 4.5 improves prompt adherence, alignment, and aesthetics over Seedream 4.0. Seedream 4 remains useful when you want the broader `n` batch workflow and a straightforward 1K-to-4K API surface.
How do I retrieve the final images?
Submit the job, store the returned `task_id`, and poll the standard task status endpoint until the task becomes `finished` or `failed`. You can also provide a `callback_url` if you want webhook notifications.
Why Use PoYo for Seedream 4 API Access
01
Instant Access
Create an API key in minutes and start testing Seedream 4 immediately. No provider waitlist, no separate approval flow.
02
Doc-Aligned Async Workflow
PoYo exposes Seedream 4 through the standard submit and status endpoints, which keeps your integration pattern consistent across image models.
03
Simple Pay-as-You-Go Pricing
Start at 5 credits per image with no subscription fee, then scale batch generation using `n` without changing your request structure.
04
One Key for More Models
Use the same PoYo account for Seedream, FLUX, GPT Image, Nano Banana, video models, and more when your product needs several creative backends.