bytedance/seedream/4.5/edit
A new-generation image creation model from ByteDance, Seedream 4.5 integrates text-to-image generation and image editing into a single unified architecture, delivering high-fidelity visuals, precise prompt control, and s...
PNG, JPEG, WebP, or GIF · 20 MiB maximum each
Example output — click Run to generate your own
Send your first request
OpenAI-compatible endpoint with unified authentication and usage tracking.
https://api.sandbase.ai/v1/runbytedance/seedream/4.5/editInput Schema
5 parameters · 2 required · 3 optional
| Parameter | Type | Required | Description |
|---|---|---|---|
images | string[] | Required | The URLs of the images to use for image-to-image generation or image editing. |
prompt | string | Required | The text prompt to generate an image from. · Min length: 3 · Max length: 50000 · Default: "Replace the clothing in image 1 with the outfit from image 2." |
image_size | object | Optional | — |
aspect_ratio | string | Optional | The aspect ratio of the generated image. · Options: 21:9, 16:9, 3:2, 4:3, 1:1, 3:4, 2:3, 9:16 21:916:93:24:31:13:42:39:16 |
output_format | string | Optional | The format of the generated image. · Options: jpeg, png · Default: "png" jpegpng |
Output Schema
| Field | Type | Description |
|---|---|---|
id | string | Unique identifier for the generation task |
status | string | Task status: pending, running, completed, failed, timeout |
model | string | Model used for the generation |
outputs | array | Array of output items |
outputs[].url | string | URL of the generated artifact |
outputs[].content_type | string | MIME type (e.g. image/png, video/mp4) |
error | object | null | Error details if failed, null on success |
error.type | string | Machine-readable error type code |
error.message | string | Human-readable error description |
Async Workflow
This model uses asynchronous execution. Submit a request and poll for the result.
- Submit — POST to /v1/run, receive an
id - Poll — GET /v1/run/{id} until status is
completed,failed, ortimeout - Retrieve — Read
outputsfrom the completed response
Code Examples
Ready-to-run snippets
# Step 1: Submit
curl -X POST https://api.sandbase.ai/v1/run \
-H "Content-Type: application/json" \
-H "Authorization: Bearer YOUR_API_KEY" \
-d '{
"model": "bytedance/seedream/4.5/edit",
"images": [
"https://p9-arcosite.byteimg.com/tos-cn-i-goo7wpa0wc/4b4464161cf3463db6f9463b10939178~tplv-goo7wpa0wc-image.image",
"https://p9-arcosite.byteimg.com/tos-cn-i-goo7wpa0wc/c23d1b0528a14cb08b684307eabdcc9b~tplv-goo7wpa0wc-image.image"
],
"prompt": "Replace the clothing in image 1 with the outfit from image 2.",
"output_format": "png"
}'
# Step 2: Poll result (replace <id>)
curl https://api.sandbase.ai/v1/run/<id> \
-H "Authorization: Bearer YOUR_API_KEY"API README
Seedream 4.5 Image Editing
Seedream 4.5 edits one or more supplied images from a written instruction. It accurately identifies the main subjects in multi-image editing. It also preserves details from the reference images.
The model strengthens typography and dense-text rendering for professional visual creatives. It supports designer-level composition and clear small text for posters and brand visuals. Across several inputs, it identifies target elements for controllable and consistent multi-image generation.
Highlights
- Reference-image consistency. Seedream 4.5 preserves facial features, lighting, color tone, and other reference details. These characteristics can remain stable while the requested edit is applied.
- Designer-level typography. The model supports professional composition and typography for posters and brand visuals. It renders small text clearly and readably within dense layouts.
- Accurate multi-image editing. Seedream 4.5 identifies target elements across multiple input images. This enables controllable and consistent multi-image generation.
- Improved model performance. Compared with Seedream 4.0, Seedream 4.5 improves prompt adherence, alignment, and aesthetics. These are the core dimensions reported in the official model evaluation.
Pricing
| Unit | Price |
|---|---|
| Per request | $0.04 |
When to Use
| ✅ Good fit | ❌ Consider alternatives |
|---|---|
| Change a material while keeping the subject's pose and the flowing shape of the original garment. | Generate from text without any reference image; choose a text-to-image model. |
| Translate text in an image while keeping all other content unchanged. | Deliver SVG, editable vector artwork, or a layered design file; choose a tool that supports those outputs. |
| Change a selected word and specify its color and type style. | Produce video or timeline animation; choose a video-generation model. |
| Build a product display with a title, centered product, price, and designed background. | Request a width or height outside 512–8,192 pixels; choose a tool that supports the target dimensions. |
| Identify target elements across multiple input images for a controlled and consistent result. | Require an output other than JPEG or PNG; choose a tool that supports the target file format. |
Prompt Guide
Assign a role to each input image, name the target of the edit, and state the details that must remain unchanged. Put exact text and placement instructions directly in the prompt when the edit includes typography.
Goal: [the requested edit]
Image roles: [what image 1, image 2, and any additional images contribute]
Target: [the person, object, garment, or text to change]
Keep: [facial features, lighting, color tone, pose, or layout to preserve]
Change: [the exact replacement or transformation]
Typography: [exact text, type style, color, and placement when needed]
Multi-image editing example using the required request fields:
{
"images": [
"<person-image-url>",
"<outfit-image-url>"
],
"prompt": "Replace the clothing in image 1 with the outfit from image 2."
}
Technical Specs
| Spec | Value |
|---|---|
| Model ID | bytedance/seedream/4.5/edit |
| Required inputs | images, prompt |
| Prompt length | 3–50,000 characters |
| Aspect ratios | 21:9, 16:9, 3:2, 4:3, 1:1, 3:4, 2:3, 9:16 |
| Custom dimensions | Width and height from 512 to 8,192 pixels |
| Output formats | jpeg, png; default png |
| Output | Image URL with optional content type |
Related
- Seedream 4.5 — Generate a new image from a text prompt when reference-image editing is not required.