
ByteDance’s newest Seedream: realistic images and consistent multi-step edits with up to 10 reference images, clean text and transparent backgrounds, at 1.5K or 2K.
A failed run is refunded automatically.
Same family, different price
Same model, same balance. Call it over the REST API or let your agent use it through MCP.
bytedance/seedream-5.5curl https://api.fcaai.ai/v1/generations \
-H "Authorization: Bearer $FCAAI_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "bytedance/seedream-5.5",
"input": {
"prompt": "Glass villa on an Aegean cliff at blue hour, travertine terrace, infinity pool catching the last light, 35 mm",
"aspect_ratio": "16:9",
"resolution": "2k"
}
}'Details
Seedream 5.5 is ByteDance’s newest image model. It creates images from a text prompt and edits or combines up to 10 reference images while keeping the parts you did not ask to change stable, so you can refine a picture over several rounds. It renders clean text and can return a transparent background.
Attach a product shot and ask to remove tags, stands or dust while keeping the exact angle, light and material details.
Describe the scene and give the headline and supporting lines in quotes with their placement.
Generate a single object on a transparent background for menus, stores or layouts.
Place the product from one image into the setting of another and match the lighting.
Seedream 5.5 takes your inputs, runs with the settings you choose and puts the result in your Library. Failed runs are refunded automatically.
What works best with Seedream 5.5.
Per run. Studio is priced in credits (1 credit = $0.01); the API column is the developer rate in US dollars, billed from the same balance. Failed runs cost nothing.
| Settings | Studio (credits) | API (USD) |
|---|---|---|
| 1K | 3 cr | $0.026 |
| 2Kdefault | 5 cr | $0.046 |
Estimates at list price; the exact price is on the Generate button before you run.
Try the idea at the lowest setting first, then run the final with the same prompt.
Same model; the developer API is billed in US dollars. POST /v1/generations with the model id below, then poll GET /v1/generations/{id}?wait=60 or use a webhook. GET /v1/models/{id} returns the price for your settings without running anything.
curl https://api.fcaai.ai/v1/generations \
-H "Authorization: Bearer $FCAAI_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "bytedance/seedream-5.5",
"input": {
"prompt": "Glass villa on an Aegean cliff at blue hour, travertine terrace, infinity pool catching the last light, 35 mm",
"aspect_ratio": "16:9",
"resolution": "2k"
}
}'| Field | Type | Values |
|---|---|---|
| promptrequired | string | ≤ 9,500 |
| images | array |
Your turn
The price is on the button before you press it; failed runs are refunded.
Prompt only. The model builds the scene from your words.
An image plus a sentence: what to change and what to keep.
Pulls characters, objects, motion or rhythm from separate files that you mention in the prompt.
Write the subject, setting, composition, lighting and style. For text in the image, give the exact words in quotes and say where they go.
Attach up to 10 images. Refer to them by order (the first image, the second image) and say what each one contributes and what must stay the same.
Pick an aspect ratio, or Auto to keep the shape of your first image when editing. Choose 1k (renders at 1.5K) or 2k.
Leave Background on Auto, or choose Transparent for cut-outs, logos and stickers (PNG with an alpha channel).
Run it, then send the result back as a reference with the next instruction to keep editing step by step.
Every setting works the same in Studio, Canvas, mobile and the API; the options come from one source.
Pick the shape for where the result will be used: 4:5 for posts, 16:9 for covers, 9:16 for stories.
Higher resolutions take longer and use more credits.
Several variations from one prompt; credits multiply with the count.
What Seedream 5.5 accepts and the limits that apply. Everything here comes from the model’s catalog entry.
| Input | Required | Limit | What it’s for |
|---|---|---|---|
| Text prompt | Yes | 9,500 characters | Describe what you want in plain words: the subject, the setting, the light and the style. |
| Source image | No | 1 | The image to work on. A sharp, well-lit picture gives the best result. |
| up to 10 |
| aspect_ratio | string | 1:1 · 4:5 · 16:9 · 9:16 · 3:2 · 21:9 · 4:3 · 3:4 · 2:3 |
| resolution | string | 1k · 2k |
| count | integer | 1–4 |
| background | string | auto · opaque · transparent |
Up to 10.
1k (rendered at 1.5K, about 2.4 MP) or 2k (about 4 MP).
Yes. Set Background to Transparent and you get a PNG with an alpha channel, at the same price.
1:1, 4:5, 16:9, 9:16, 3:2, 21:9, 4:3, 3:4 or 2:3, or Auto when editing.
The price depends on the settings you pick and is shown on the Generate button before you run. The pricing table below lists typical image prices for Studio in credits (1 credit = $0.01) and for the API in US dollars.
You are not charged. The reserved amount returns to your balance automatically.
Yes. Use the model id shown in the API section with POST /v1/generations. Add "dry_run": true to get a price quote without running.
Yes, on every paid plan. The Free plan is for personal use.
Yes. Seedream 5.5 is available as a node in Canvas and in every template that uses it.