Image GenerationActive
ByteDance / seedream-v5-pro
seedream-v5-pro
ByteDance Seedream V5 Pro generates images from text and edits a base image using up to 10 references. It’s tuned for infographics, layouts, and clear typography.
Text to ImageImage to Image
Model ID
seedream-v5-pro
Provider
bytedance
Updated
1784038055
wiro playground—bytedance/seedream-v5-pro
Updated 1784038055
Overview
Seedream V5 Pro is ByteDance’s Seedream 5.0 Pro image model. It generates a single image from a written brief. It can also edit an existing image using reference images. The model focuses on layout planning, typography, and region-aware edits. This helps you ship posters, UI concepts, and infographics that stay readable.
What you can build
- Product and brand visuals, like hero shots, ads, and campaign key art
- Dense infographics with timelines, charts, labels, and small text blocks
- Poster designs with readable headlines and structured layout
- UI-style mockups for homepages, app screens, and storefront banners
- Controlled photo edits, like object swaps, recolors, and material changes
- Multi-reference composites that merge elements from several images
- Multilingual typography graphics for menus, packaging, and signage
Inputs
- A required written description of what to generate, or what to change in the image.
- Optional reference images provided as public URLs or uploaded files. Use JPG or PNG. You can supply up to 10 images.
- A resolution tier choice in the 1K or 2K range. This sets the output pixel budget.
- An optional aspect ratio choice, such as square, portrait, landscape, or wide.
- A choice of output file type. Pick JPEG or PNG.
- An option to keep or remove a visible watermark in the final image.
Outputs
- One generated image, delivered as a downloadable JPEG or PNG.
- A hosted download link for the image file.
- Image metadata when available, such as MIME type, file name, file size, and pixel dimensions.
Limitations
- Output resolution is limited to the 1K–2K class. Very wide ratios can reduce fine detail.
- Image editing depends on clear references. If you supply more than 10, the extra images won’t apply.
- Long paragraphs and tiny text can still contain typos or spacing issues.
- Precise edits work best with clear visual cues. Unmarked images make local edits harder.
- Low-quality inputs can cause artifacts. Watch for blur, heavy JPEG compression, and existing watermarks.
Safety & compliance
- The serving stack can apply safety filtering. Some prompts or images may be blocked.
- Only upload images you own or have rights to use.
- Get consent for private people and faces. Don’t impersonate real individuals.
- Don’t generate illegal, hateful, or explicit sexual content. Follow your org’s policies.
Example prompts
Great starting points for seedream-v5-pro.
Keep the exact same man, same face and beard. Place him leaning against a neon-lit rainy Tokyo street at night, reflections on wet pavement, moody cyberpunk color grading, shallow depth of field, cinematic photography.Image Generation
Keep the same valley and river. Reimagine it at dawn with thick rolling fog in the low areas, dramatic misty light rays through the mountains, moody muted colors, serene atmospheric landscape photography.Image Generation
Keep the same burger, same ingredients and plating. Restage it on a rustic wooden table in a warm cozy diner, soft window light, fries and a drink beside it, shallow depth of field, mouth-watering food photography.Image Generation
Keep the same skyline and river. Reimagine it at night with glowing skyscraper lights reflecting on the water, deep blue twilight sky, a few drifting clouds, long-exposure cityscape photography.Image Generation
API quick start
Run seedream-v5-pro with a single API call.
POST https://api.wiro.ai/v1/Run/ByteDance/seedream-v5-pro
{
"prompt": "Keep the same valley and river. Reimagine…",
"resolution": "1k",
"watermark": "false",
"inputImage": "https://your-cdn.com/input.png"
}