Introduction
Seedream 4.5 Image Editing changes, restyles or combines pictures from a prompt on Seedream 4.5, ByteDance's image model. It takes one to fourteen references, which you name Image 1, Image 2 and so on, and returns one image at 2K or 4K. The platform adds no content filter. For a new picture from words alone, use Seedream 4.5 Text to image.
Choose a model by what you need
| What you need | Model |
|---|---|
| a new image from words alone | Seedream 4.5 Text to image |
| edit, restyle or combine up to 14 pictures | Seedream 4.5 Image Editing (you are here) |
Key features & highlights
Up to fourteen references
A woman from one picture and a gown from another came together in one image in our tests. ByteDance says 4.5 recognises the main subject more accurately in multi-image edits.
Keeps her recognisable
In our tests the face, freckles and red hair carried over to a new outfit and a new place. Under a heavy restyle such as black and white, features shifted slightly.
Flat price
$0.04 per image whatever the number of references, 2k or 4k.
Unfiltered
Nothing is refused or filtered by the platform. In our tests Strong and Very strong came back toned down, and Mild, Moderate and Maximum came back as written.
Request parameters
| Parameter | Required | Description |
|---|---|---|
prompt | Yes | 1 to 3,000 characters; required. Name each picture as Image 1, Image 2 and so on, in the order you list them, and say what to keep and what to change. at least 1 characters · at most 3000 characters |
image_urls | Yes | 1 to 14 pictures; required. JPEG, PNG or WebP, up to 30 MB each, as publicly reachable URLs or platform uploads. at least 1 items · at most 14 items |
aspect_ratio | No | One of 1:1 (default), 4:3, 3:4, 16:9, 9:16, 3:2, 2:3 or 21:9. There is no auto, so pick the ratio closest to your source picture. |
resolution | No | 2k (about 4 megapixels, default) or 4k (about 16). Both cost the same. |
Pricing
$0.04 / image
$0.04 per image. The number of reference images, the resolution (2k or 4k) and the aspect ratio do not change the price. Failed requests are not charged.
Use cases
Outfit swaps
Take a garment from one picture and put it on the person in another.
New scenes for one character
Move the same portrait to a terrace, a staircase or a street at another time of day.
Restyling
Turn a portrait into an era, a film look or a lighting mood.
Prompt guide & pro tips
Give every picture a job
Write "the woman from Image 1" and "the gown from Image 2". References are numbered in the order you list them.
Say what to keep
Name what must stay: her face, freckles and long red hair. Then give the change: the place, the outfit and the light.
Pick the closest ratio
There is no auto ratio. For a portrait, 3:4 or 2:3 keeps its shape; for a landscape, 4:3 or 3:2.
Do not rely on exact pixels
At 2k the output is about 4 megapixels and follows the source a little: the same 3:4 request came back at 1776 x 2352.
Expect a new image every time
There is no seed. Save results you like.
API quick start
# Generate the idempotency key once; reuse it when retrying after a lost response.
export SPICY_IDEMPOTENCY_KEY="${SPICY_IDEMPOTENCY_KEY:-$(uuidgen)}"
curl -X POST https://api.spicyapi.ai/api/v1/jobs/createTask \
-H "Authorization: Bearer $SPICY_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: $SPICY_IDEMPOTENCY_KEY" \
-d '{
"model": "bytedance/seedream-4.5/edit",
"input": {
"aspect_ratio": "1:1",
"image_urls": [
"https://example.com/reference.png"
],
"prompt": "A ceramic teapot on a linen cloth, soft window light, muted palette",
"resolution": "2k"
}
}'
# Copy data.taskId from the response into TASK_ID, then poll the status.
TASK_ID="TASK_ID_FROM_RESPONSE"
curl "https://api.spicyapi.ai/api/v1/jobs/recordInfo?taskId=$TASK_ID" \
-H "Authorization: Bearer $SPICY_API_KEY"
See the API reference for authentication, the complete schema, polling, and error handling.
