This is the one-page reference for prompting Seedream — ByteDance's flagship text-to-image and image-editing model. Treat Seedream 4.5 as the current default; Seedream 5.0 Pro (released July 8, 2026) is the newest. Unlike Midjourney, Seedream takes no --parameters: it reads full descriptive sentences, so you write everything in plain English and brief it like a photographer, not with comma-separated tags.

New to it? Start with the best Seedream prompts roundup, then keep this sheet open while you build your own. For a deeper walk-through of skin and lighting, see how to prompt Seedream for photorealism.

Advertisement

The prompt formula

A strong Seedream prompt is one flowing natural-language paragraph, not a tag list. Word order matters — elements earlier in the prompt get priority, so lead with the most important subject or detail. Aim for the 30–100 word sweet spot: under about 15 words leaves too much to interpretation, and over about 150 stacks conflicting instructions.

The structure that works: Subject + scene/action + composition/framing + lighting + style + [text in quotes if any] + aspect ratio + resolution.

Skeleton to copy:

[Subject, described concretely] [doing what], [composition/framing],
in [setting/location with real detail], lit with [lighting setup],
[style/medium]. [Any on-image text in "double quotes", font named.]
[X:Y] aspect ratio. Render in [2K/4K].

The same skeleton, filled in:

A weathered fisherman in his sixties wearing a knitted grey sweater, gazing
just off-camera, tight head-and-shoulders portrait, on a misty harbour wall
at dawn, lit with soft golden-hour backlighting and a gentle rim light,
photorealistic editorial style with natural skin texture and shallow depth
of field. 4:5 aspect ratio. Render in 4K.

Why it works: the most important subject leads, every factor is present and specific, and it reads as a scene a photographer could shoot — which is exactly what Seedream was trained to reward. It lands inside the 30–100 word window with no conflicting instructions.

Aspect ratios & resolution

State the ratio in words ("9:16 aspect ratio" or "1:1 square"); there are no flags. Put it near the end of the prompt with the resolution. These are the ratios that render reliably.

Aspect ratioShapeTypical use
1:1SquareProfile pics, album art, Instagram grid, product tiles
4:5PortraitInstagram feed portrait — the max vertical the feed allows
3:4PortraitPortraits, mobile-first cards, catalogue shots
2:3PortraitBook covers, posters, Pinterest pins
3:2LandscapeClassic photography, editorial, prints
4:3LandscapePresentation slides, classic monitors, blog headers
16:9WideYouTube thumbnails, hero banners, desktop wallpaper
9:16Tall portraitStories, Reels, TikTok, phone wallpapers
21:9Ultra-wideCinematic scenes, ultrawide monitors, website headers

2K vs 4K: Seedream 4.5 outputs up to roughly 2048px / 4K. Ask for "2K" for screen and social work and "4K" for print or anything you will crop into. Higher resolution also renders cleaner typography, so choose 4K whenever the image contains words.

Lighting modifiers

Lighting sets the mood faster than any other factor. Name a real setup and Seedream renders it convincingly. Drop one phrase into the lighting slot of the formula.

ModifierEffect
Three-point softbox setupClean, even, flattering studio light — the safe default for portraits and products
Golden-hour backlightingWarm glow and long shadows, halo around edges — dreamy and cinematic
Soft overcast diffusionFlat, shadowless, gentle light — natural skin and honest product colour
Rim lightBright edge that separates the subject from a dark background
Rembrandt lightingSmall triangle of light on the shadowed cheek — classic portrait look
Chiaroscuro, high contrastDramatic light-and-shadow, deep blacks — painterly and moody
High-key lightingBright, airy, low-shadow — beauty, e-commerce, clean and optimistic
Low-key lightingMostly dark with selective highlights — noir, luxury, tension
Neon / practical lightsColoured urban glow (magenta, cyan) reflecting on wet surfaces
Hard direct sunlightCrisp, high-contrast shadows — bold fashion and street work
Candlelight / firelightWarm, flickering, low and intimate — cosy and romantic scenes

Camera & lens modifiers

Real lens language tells Seedream how the scene should compress, blur, and frame. Fold these into the sentence ("shot on 85mm f/1.4 with shallow depth of field").

ModifierEffect
85mm f/1.4Flattering portrait compression with creamy background blur
35mmNatural, documentary, street perspective close to how the eye sees
24mm wide-angleExpansive scenes, interiors, landscapes — some edge stretch
100mm macroExtreme close-up detail — jewellery, food, textures, insects
Shallow depth of fieldSharp subject, soft background — isolates and draws the eye
Deep depth of fieldEverything sharp front to back — landscapes, architecture
BokehSoft round out-of-focus highlights behind the subject
Tilt-shiftSelective focus band that makes real scenes look miniature
Drone / aerial shotTop-down or high overhead view for scale and pattern
Low angle / worm's-eyeLooking up — makes the subject tower and feel heroic
FisheyeStrong curved distortion for a playful, immersive look
Advertisement

Style & medium modifiers

The style factor decides whether you get a photo, a render, or an illustration. Name one clearly; mixing three fights the model.

ModifierEffect
PhotorealisticLooks like a real photograph — natural texture, real lighting physics
CinematicFilm colour grade, wide-frame drama, moody contrast and atmosphere
3D render (Octane / Blender)Clean CGI with glossy materials and studio reflections
WatercolourSoft washes, bleeding edges, visible paper texture
IsometricAngled 3D-tile look for scenes, rooms, and game-style diagrams
Flat vector illustrationBold shapes, limited palette, no gradients — icons and web art
Film grain, 35mm analogGrainy, slightly faded, nostalgic film-camera aesthetic
Product studioSeamless backdrop, crisp reflections, e-commerce-ready lighting
Anime / mangaCel-shaded characters with expressive lines and flat colour
Line art / sketchPen or pencil linework, minimal or no colour
Oil paintingVisible brushstrokes, rich impasto texture, classical feel

Text-rendering rules

Seedream has best-in-class text rendering for an image model, in English and Chinese — rare and reliable enough to build posters, product key visuals, and social ads around. Follow these rules and it renders words cleanly. For ready-made poster skeletons, see the best Seedream prompts.

RuleWhy
Wrap the exact words in "double quotes"Tells the model precisely which characters to render — no paraphrasing
Keep it to 3–5 wordsShort strings (4–10 letters) render cleanest; long lines garble
Name the font style"bold sans-serif", "elegant serif", "handwritten script", or "grotesk" guides the letterforms
Describe placement"headline top-center, subtitle below, CTA bottom-right" controls the layout
Render at higher resolution4K produces crisper, more legible type than 2K
English and Chinese both workSeedream renders both scripts reliably — state the language if it matters

Multi-image & editing rules

Multi-image work is Seedream's standout. It accepts text, one image, or up to ~6 reference images — enabling multi-image fusion, single-image edits, and sequential sets that stay on-model.

  • Reference order: feed them identity first, then style, then palette, then material/layout.
  • Label each reference: spell out its role — "use image 1 for the product, image 2 for the background style, image 3 for the colour palette."
  • Editing pattern — Action + Object + Attributes: say what to change and what to keep identical: "keep pose, lighting and background the same; change only the jacket to red."
  • Consistent characters: supplying a reference photo preserves faces and clothing across a whole set.
  • Emphasis weighting: boost an element with (worn leather satchel:1.2).
Use image 1 for the product and image 2 for the background style. Place the
sneaker on the marble surface from image 2, keep the shoe's shape, colour and
logo exactly the same, match the lighting to the background. Emphasise the
(glossy sole:1.2). 4:5 aspect ratio. Render in 4K.

Best for: composited product shots and character sets — label every reference and state what must stay identical.

Copy-paste example prompts

Five complete natural-language prompts that assemble the modifiers above. Paste any into Seedream and tweak the specifics. For more, browse the best Seedream prompts.

1. Product hero shot

A matte-black ceramic coffee mug with steam rising, three-quarter view
centered on a wet slate surface, in a dim minimalist studio, lit with a
three-point softbox setup and a subtle rim light, photorealistic product
studio style with crisp reflections and a seamless charcoal backdrop, shot
on 100mm macro with shallow depth of field. 1:1 square. Render in 4K.

Best for: e-commerce listings and ads — clean studio light and a macro lens make the material read as premium.

2. Editorial portrait

A confident woman in her thirties with short curls and a tailored linen
blazer, looking directly at camera, tight head-and-shoulders portrait, in a
sunlit loft with a soft blurred window behind her, lit with Rembrandt
lighting and soft overcast fill, photorealistic editorial style with natural
skin texture, shot on 85mm f/1.4 with shallow depth of field. 4:5 aspect
ratio. Render in 4K.

Best for: LinkedIn headshots and magazine-style portraits. See the photorealism guide for skin and lighting deep-dives.

3. Cinematic scene

A lone figure in a long coat, mid-stride at a rain-slicked neon crossroads
in a futuristic Tokyo alley at night, wide establishing shot from a low
angle, lit by magenta and cyan neon reflecting off wet asphalt, cinematic
film-grade colour with atmospheric haze, shot on 35mm with deep depth of
field. 21:9 aspect ratio. Render in 4K.

Why it works: the 21:9 ratio and neon lighting do the cinematic heavy lifting, and the subject leads so word order keeps the figure the focus.

4. Flat illustration

A friendly flat vector illustration of a person watering a large houseplant
in a cozy apartment, centered composition, warm limited palette of terracotta,
sage and cream, bold clean shapes with no gradients, soft even lighting,
modern flat illustration style for a web hero. 16:9 aspect ratio. Render in 2K.

Best for: landing-page graphics and blog headers where a clean, on-brand illustration beats a photo.

5. Text poster

A retro travel poster with the headline "VISIT MARS" in a bold geometric
sans-serif across the top and the subtitle "Red Planet Awaits" below it. A
stylised red desert landscape with a distant domed colony under a pink sky,
mid-century illustration style with a warm limited palette, headline
top-center and subtitle beneath. 2:3 aspect ratio. Render in 4K.

Why it works: the exact text sits in quotes and stays short, the font style and placement are named, and 4K keeps the lettering crisp.

Frequently Asked Questions

Does Seedream accept --parameters like Midjourney?

No. Seedream has no --ar, --v, or --style flags. It reads full descriptive sentences, so you state everything in plain English instead: "9:16 aspect ratio", "soft golden-hour backlighting", "render in 4K". Writing comma-separated tag-soup or Midjourney flags actually hurts results — brief it like a photographer, in prose.

What is the maximum resolution Seedream can output?

Seedream 4.5 outputs up to roughly 2048x2048, or 4K. It does not always default to the maximum, so add "render in 2K" for screen and social work or "render in 4K" for print and fine detail. Higher resolution also renders typography more cleanly, so ask for 4K whenever the image contains words.

How do I set the aspect ratio in Seedream?

State it in words inside the prompt — for example, "9:16 aspect ratio" or "1:1 square". There are no flags. Common ratios that work well are 1:1, 4:5, 3:4, 2:3, 3:2, 4:3, 16:9, 9:16, and 21:9. Put the ratio near the end of the prompt alongside the resolution.

Which Seedream model is best for text in images?

Seedream 4.5 already has best-in-class text rendering for both English and Chinese, and Seedream 5.0 Pro (released July 8, 2026) is the newest and reasons harder about layout. For posters and product key visuals, use either, keep the text to 3-5 words in double quotes, name the font style, and render in 4K for the crispest letterforms.

How many reference images can Seedream use in one prompt?

Up to about six reference images in a single prompt. Feed it identity, style, palette, and material references, and label what each one is for — for example, "use image 1 for the product, image 2 for the background style, image 3 for the colour palette." This is how you keep characters and products on-model across a set.

How do I edit an image instead of regenerating it?

Use Seedream's Edit mode and describe only the one change plus what must stay identical — the Action + Object + Attributes pattern: "keep the pose, lighting, and background exactly the same; change only the jacket to red." Supplying a reference photo preserves faces and clothing. Regenerating from scratch loses the parts you already liked.

How long should a Seedream prompt be?

Aim for 30 to 100 words. Under about 15 words leaves too much to interpretation, and over about 150 stacks conflicting instructions that fight each other. Lead with the most important subject or detail, because word order matters — elements earlier in the prompt get priority.

Where can I use Seedream?

Seedream is ByteDance's flagship text-to-image and image-editing model, with 4.0, 4.5, and the newest 5.0 Pro in the lineup. It is available through ByteDance and BytePlus ModelArk, the Doubao, Dreamina, and Jimeng apps, and third-party platforms including fal.ai, WaveSpeed, Replicate, Picsart, and Freepik.

Advertisement