WaveSpeed.ai
Home/Explore/Seedream AI Models/bytedance/seedream-v4.5/sequential
text-to-image

text-to-image

ByteDance Seedream 4.5 Sequential

bytedance/seedream-v4.5/sequential

Seedream 4.5 Sequential generates multi-image sets with consistent characters and objects, unifying palette, lighting, and style across all outputs. Supports up to 4K results for campaigns, storyboards, and product lines. Ready-to-use REST inference API, best performance, no cold starts, affordable pricing.

Input
width
height
If enabled, the output will be encoded into a BASE64 string instead of a URL. This property is only available through the API.
If set to true, the function will wait for the result to be generated and uploaded before returning the response. It allows you to get the result directly in the response. This property is only available through the API.

Idle

Generate a sequence of 6 atmospheric city alley scenes that all take place in the same narrow street, with consistent architecture, signs, perspective and overall style. Style: realistic photography with a slight cinematic film look, soft grain, subtle lens bloom, inspired by Japanese backstreets. Scene 1 – Early Morning: the alley is quiet and almost empty, cool blue light, shutters half-closed, a faint mist on the ground. Scene 2 – Late Morning: shops are open, warmer sunlight enters the alley, a few people walking and bicycles parked, brighter colors. Scene 3 – Golden Hour: warm orange light from the low sun at the end of the alley, long shadows, vending machines glowing softly. Scene 4 – Blue Hour: sky is deep blue, shop signs and lanterns turn on, reflections on slightly wet pavement. Keep the street layout, buildings, signs and vantage point consistent across all 6 images so they feel like different times of day in the exact same location
Generate a sequence of 6 atmospheric city alley scenes that all take place in the same narrow street, with consistent architecture, signs, perspective and overall style. Style: realistic photography with a slight cinematic film look, soft grain, subtle lens bloom, inspired by Japanese backstreets. Scene 1 – Early Morning: the alley is quiet and almost empty, cool blue light, shutters half-closed, a faint mist on the ground. Scene 2 – Late Morning: shops are open, warmer sunlight enters the alley, a few people walking and bicycles parked, brighter colors. Scene 3 – Golden Hour: warm orange light from the low sun at the end of the alley, long shadows, vending machines glowing softly. Scene 4 – Blue Hour: sky is deep blue, shop signs and lanterns turn on, reflections on slightly wet pavement. Keep the street layout, buildings, signs and vantage point consistent across all 6 images so they feel like different times of day in the exact same location
Generate a sequence of 6 atmospheric city alley scenes that all take place in the same narrow street, with consistent architecture, signs, perspective and overall style. Style: realistic photography with a slight cinematic film look, soft grain, subtle lens bloom, inspired by Japanese backstreets. Scene 1 – Early Morning: the alley is quiet and almost empty, cool blue light, shutters half-closed, a faint mist on the ground. Scene 2 – Late Morning: shops are open, warmer sunlight enters the alley, a few people walking and bicycles parked, brighter colors. Scene 3 – Golden Hour: warm orange light from the low sun at the end of the alley, long shadows, vending machines glowing softly. Scene 4 – Blue Hour: sky is deep blue, shop signs and lanterns turn on, reflections on slightly wet pavement. Keep the street layout, buildings, signs and vantage point consistent across all 6 images so they feel like different times of day in the exact same location
Generate a sequence of 6 atmospheric city alley scenes that all take place in the same narrow street, with consistent architecture, signs, perspective and overall style. Style: realistic photography with a slight cinematic film look, soft grain, subtle lens bloom, inspired by Japanese backstreets. Scene 1 – Early Morning: the alley is quiet and almost empty, cool blue light, shutters half-closed, a faint mist on the ground. Scene 2 – Late Morning: shops are open, warmer sunlight enters the alley, a few people walking and bicycles parked, brighter colors. Scene 3 – Golden Hour: warm orange light from the low sun at the end of the alley, long shadows, vending machines glowing softly. Scene 4 – Blue Hour: sky is deep blue, shop signs and lanterns turn on, reflections on slightly wet pavement. Keep the street layout, buildings, signs and vantage point consistent across all 6 images so they feel like different times of day in the exact same location

Your request will cost $0.04 per run.

For $1 you can run this model approximately 25 times.

One more thing::

ExamplesView all

Generate a sequence of 6 atmospheric city alley scenes that all take place in the same narrow street, with consistent architecture, signs, perspective and overall style. Style: realistic photography with a slight cinematic film look, soft grain, subtle lens bloom, inspired by Japanese backstreets. Scene 1 – Early Morning: the alley is quiet and almost empty, cool blue light, shutters half-closed, a faint mist on the ground. Scene 2 – Late Morning: shops are open, warmer sunlight enters the alley, a few people walking and bicycles parked, brighter colors. Scene 3 – Golden Hour: warm orange light from the low sun at the end of the alley, long shadows, vending machines glowing softly. Scene 4 – Blue Hour: sky is deep blue, shop signs and lanterns turn on, reflections on slightly wet pavement. Keep the street layout, buildings, signs and vantage point consistent across all 6 images so they feel like different times of day in the exact same location
Generate a set of 3 consecutive illustrations in a gritty action-anime style: first, a lone rider speeds through a shattered desert highway on a futuristic motorcycle, pursued by armored raiders. The next illustration shows an intense mid-air leap over a collapsed bridge as explosions erupt behind them. The final scene captures the rider skidding to safety atop a cliff, raising dust while the burning wreckage of the pursuers lights up the wasteland below.

README

bytedance/seedream-v4.5/sequential

Seedream 4.5 Sequential is ByteDance’s multi-image generation model for creating whole series of images in one go. It keeps characters, props, and style consistent across all outputs, making it ideal for KVs, comic panels, and any visual set that should “look like one universe”.

Model highlights

  • Character consistency – Locks onto the same character identity (face, hairstyle, body shape) across all generated frames.
  • Object & prop stability – Reuses key objects (products, logos, props) so they don’t randomly change between images.
  • Unified visual style – Maintains palette, lighting, camera feel, and rendering style across the whole set.
  • Multi-image output – Generate several images in a single request, all driven by the same prompt.
  • 4K-ready detail – Supports resolutions up to 4096 × 4096 per image for hero KVs and print-adjacent work.
  • Typography aware – Strong on-image text rendering for branded content, titles, and UI-like elements.

Best suited for

  • Comic strips and story panels with recurring characters
  • Brand KV series and campaign sets built around the same hero figure
  • Product lineup / colourway visualisation
  • Storyboard or animatic keyframes
  • Social media content series that must feel coherent in the grid
  • Multi-step marketing journeys (awareness → consideration → conversion visuals)

Pricing

Billing is per generated image, controlled by max_images.

  • $0.04 per image
  • Formula: total_price = $0.04 × max_images

Example costs:

max_imagesTotal price
1$0.04
4$0.16
8$0.32

The exact price for your chosen settings is always shown in the WaveSpeedAI interface before you run the job.

How to use

  1. Enter your prompt Describe the scene, characters, and what must stay consistent, e.g. “Same girl with red hoodie and headphones, different city locations, cinematic lighting.”

  2. Set max_images Choose how many images you want in the series. Each one will follow the same prompt and consistency logic.

  3. Set size Choose width and height, up to 4096 × 4096 for maximum detail.

  4. Run and review Generate the series, check continuity across faces, outfits, and props, then refine the prompt for another pass if needed.

Notes

  • Please set the max_image first, and then input how many images you want to generate in prompt! Such as:

    • max_image = 4.

    Prompt: I want to generate 4 images... + (your prompt)

Model comparison & related tools

  • Nano Banana Pro Google’s ultra-low-cost, high-throughput T2I model. Great for lots of individual images or quick ideation, but it doesn’t provide built-in cross-image character locking.

  • Seedream V4 – sequential ByteDance’s single-image Seedream generator with rich detail and stylish output. Ideal when you want one-off hero shots or standalone illustrations rather than a strictly consistent series.

  • Qwen Image Edit Plus Qwen-Image is a 20B MMDiT-based text-to-image generation model, especially strong at native text rendering in both English and Chinese. It is a powerful creative tool for posters, comics, and visual storytelling, while also excelling at general image generation from photorealism to anime.