Seedream 4.5 NSFW Text to Image API: 2K, 4K

bytedance/seedream-4.5/text-to-image

Text to image on Seedream 4.5, ByteDance's image model. Prompts up to 3,000 characters, eight aspect ratios, 2K or 4K, text that came out as quoted in our tests, one flat price per image, and no content filter.

From $0.04/image

Run model View API

Input

Text to image

Prompt431/3000

Aspect ratio

Resolution

Safety checkerOff

SpicyAPI does not run a safety checker on this model. Your prompt and media go to the model as written, and the model itself decides what comes back.

Checking your account before enabling submission and pricing.

Output

Ready for a new run

Example result

Edit

Split into layers

Animate

{
  "model": "bytedance/seedream-4.5/text-to-image",
  "input": {
    "aspect_ratio": "21:9",
    "resolution": "4k",
    "prompt": "Cinematic film still, ultra-wide frame. A woman in her late twenties with dark curly hair, in a backless emerald satin evening gown, leans on the rail of a Manhattan rooftop at blue hour and glances back over her shoulder at the camera. City lights stretch out below, warm string lights glow behind her, a half-empty glass of champagne rests on the ledge. 50mm, shallow depth of field, rich teal and amber palette, fine film grain."
  }
}

Introduction

Seedream 4.5 Text to image draws one image from a written prompt on Seedream 4.5, ByteDance's image model, at 2K or 4K in eight aspect ratios. In our tests it set quoted text into posters and product ads as written, and the platform adds no content filter, so NSFW prompts are not refused. For changing or combining pictures, use Seedream 4.5 Image Editing.

Choose a model by what you need

What you need	Model
a new image from words alone	Seedream 4.5 Text to image (you are here)
edit, restyle or combine up to 14 pictures	Seedream 4.5 Image Editing

Key features & highlights

Text that comes out as quoted

A film poster with its title and bottom line and a perfume bottle with its engraved name were spelled correctly in our tests. ByteDance says text rendering is better in 4.5. We only tried English.

2K and 4K at one price

About 4 megapixels at 2k and about 16 at 4k. A 21:9 frame at 4k came back at 6240 x 2656, and both resolutions cost $0.04.

Eight ratios

From 21:9 to 9:16. At 2k, 16:9 is 2848 x 1600, 3:2 is 2496 x 1664 and 1:1 is 2048 x 2048.

Unfiltered

Nothing is refused or filtered by the platform. In our tests Mild, Moderate, Strong and Maximum came back as written, and Very strong came back toned down.

Request parameters

Request parameters for Seedream 4.5 · Text to Image
Parameter	Required	Description
`prompt`	Yes	1 to 3,000 characters; required. Put each piece of text in quotes and say where it goes. at least 1 characters · at most 3000 characters
`aspect_ratio`	No	One of 1:1 (default), 4:3, 3:4, 16:9, 9:16, 3:2, 2:3 or 21:9. At 2k, 16:9 is 2848 x 1600 and 4:3 is 2304 x 1728. `1:1` · `4:3` · `3:4` · `16:9` · `9:16` · `3:2` · `2:3` · `21:9` · Default: `1:1`
`resolution`	No	2k (about 4 megapixels, default) or 4k (about 16): the widest ratio at 4k came back at 6240 x 2656. Both cost the same. `2k` · `4k` · Default: `2k`

Pricing

$0.04 / image

$0.04 per image. The resolution (2k or 4k) and the aspect ratio do not change the price. Failed requests are not charged.

Use cases

Posters and ads

A title, a tagline and a product name in one picture. Quote each string and place it.

Cinematic frames

Wide 21:9 and 16:9 stills with a lens, a palette and a light source named.

Group scenes and portraits

Four friends at a pool, a couple dancing close, a boxer in an old gym: several people with their own outfits in one frame.

Explore key use-case tutorials

Prompt guide & pro tips

Quote the text, then place it

Write "NOIR VELOURS" in quotes, then say where it goes and how it looks: engraved in gold on the bottle. Keep the strings short enough to read at the size you choose.

Name the light and the lens

A single beam of warm gold light from the left, 50mm, shallow depth of field, film grain. These details shape the result more than adjectives do.

Give each person their own sentence

For group scenes, describe each person by outfit and action in turn, as in the four-friends pool prompt.

Plan for different results

There is no seed, so the same prompt returns a different image each time. Save the ones you like and keep the prompt with them.

Describe what you want

There is no negative prompt. Say what is in the frame, not what to leave out.

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/text-to-image",
  "input": {
    "prompt": "A ceramic teapot on a linen cloth, soft window light, muted palette",
    "resolution": "2k",
    "aspect_ratio": "1:1"
  }
}'

# 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.

Missing the model you need?

Tell us what to add and where you would use it. Requests go to the people who decide what gets onboarded next.

Content test

Is Seedream 4.5 uncensored? What we tested

We ran each level on our live setup and checked every clip frame by frame. A pass means the content was actually on screen, not just that the request went through.

Content levels tested on Seedream 4.5, by mode
Level	Text to image
MildThe lightest prompts in the test	RendersTested Oct 21, 2026
ModerateWhere filtered models start to refuse	RendersTested Oct 21, 2026
StrongThe middle of the scale	RendersTested Oct 21, 2026
Very strongWhere most models start to hold back	Not tested yet
MaximumThe top of the scale	RendersTested Oct 21, 2026

Renders: on screen as written. Softened: generated, but milder than what was written. Refused: rejected. Not tested yet: we have not run that level on that mode.

How each mode behaves

All modes: Text in the picture was right in our tests, but they were English only and a few strings; check each result. There is no seed and no negative prompt, no output format switch, and each request returns one image. Identical requests return different images.
All modes: Content levels are not all equal. In our tests Very strong came back toned down in both endpoints, and in editing Strong did too. Most requests are usually done in under half a minute, longer for 4K or several references.
All modes: Every person you describe must be an adult. You are responsible for the lawfulness of what you generate and publish. Generated media is retained for 14 days and then deleted.

Prompts

Seedream 4.5 prompt examples you can copy

Paste one into the playground as it is, then change one thing at a time.

Text to image
Rooftop at blue hour
Cinematic film still, ultra-wide frame. A woman in her late twenties with dark curly hair, in a backless emerald satin evening gown, leans on the rail of a Manhattan rooftop at blue hour and glances back over her shoulder at the camera. City lights stretch out below, warm string lights glow behind her, a half-empty glass of champagne rests on the ledge. 50mm, shallow depth of field, rich teal and amber palette, fine film grain.
Why it works. A named frame, a lens, a palette and a few props give a cinematic result, and the ultra-wide ratio at 4k keeps the city readable.
Try this prompt
Text to image
1970s film poster
A 1970s-style film poster. At the top, in large cream serif capitals: "THE VELVET HOUR". In the middle, a woman in a red satin slip dress and a man in an open-collar white shirt dance close together under a single spotlight in an empty ballroom, her cheek near his. At the bottom, in small spaced capitals: "IN CINEMAS THIS WINTER". Grainy offset-print texture, burgundy and gold palette, thin gold border.
Instead of “A vintage film poster called The Velvet Hour with a dancing couple.”
Why it works. Each string of text is quoted and given a place and a type style, so the title and the bottom line come out as written.
Try this prompt
Text to image
Perfume bottle ad
Luxury perfume advertisement. A faceted black glass bottle stands on wet black stone, a single drop running down the glass, one narrow beam of warm gold light from the left. Engraved on the bottle in gold: "NOIR VELOURS". Bottom centre, in thin gold sans-serif capitals: "EAU DE PARFUM". Studio product photography, deep shadows, crisp reflections.
Why it works. The bottle, the stone and the light are described once, and the two text lines say which is engraved and which sits at the bottom.
Try this prompt

Why SpicyAPI

Why run Seedream 4.5 on SpicyAPI

No filter from us
We add no content filter of our own. What the model renders is what you get back.
Tested, not guessed
Content labels come from real generations we ran and checked, with the test date on the page.
Pay per result, in dollars
No subscription and no credits. Failed or timed-out tasks are usually not charged. Some models charge for content refusals as stated on their model page; check those billing terms before submitting.
One key for every model
One API key and balance cover image, video, audio and chat models, so you can explore more models from the same account.

About Seedream 4.5

Seedream 4.5: ByteDance's image model for 2K and 4K pictures, exact text in posters and ads, and editing with up to 14 references. $0.04 an image, and nothing is filtered.

Read the full Seedream 4.5 guide: tests, examples and FAQ