Seedance v2.5 Flash Video Generator
ByteDance Seedance v2.5 Flash generates 4–30s videos from text or a first frame, with optional synced audio, at 480p or 720p.
Overview
Seedance v2.5 Flash by ByteDance generates short videos from text, images, or both. You describe the shot, then the model produces a 4–30 second clip with coherent motion and camera movement. You can anchor the result with a first frame, and optionally guide the end pose with a last frame. It can also generate synchronized audio, so you get one file you can review and cut into an edit.
What you can build
- Storyboard shots from a written scene description
- Image-to-video animations for concept art, posters, and keyframes
- Product clips from a packshot or hero frame
- Fast variants of the same shot for A/B testing creatives
- Short cinematic beats with camera cues, blocking, and sound design notes
Inputs
- A video prompt written in natural language. Include subject, action, camera, lighting, and pacing. You can also describe sound when audio is enabled.
- An optional first-frame image to animate into a shot. Use JPG, PNG, WebP, GIF, HEIC, or HEIF. Keep it under 30 MB, 300–6000 px per side, and within a 0.4–2.5 aspect ratio.
- Optional reference images to keep a character, product, outfit, or location consistent. You can add up to 10 images. Mention them in your prompt by index, like “Image 1”.
- An optional last-frame image to steer the final pose or composition. This works only when you also provide a first frame.
- A target output resolution selection of 480p or 720p.
- A target duration from 4 to 30 seconds.
- A toggle to include synchronized audio with the visuals, or return a silent clip.
- An aspect ratio choice. If you provide a first frame, the output keeps that image’s aspect ratio.
- An output container choice of MP4 or MOV.
- A watermark toggle to add an “AI Generated” mark in the lower-right corner.
- An optional prompt rewrite step that can remove brand names and reduce filter-triggering phrasing before generation.
Outputs
The model returns a generated video file as MP4 or MOV. The clip contains the rendered frames at the requested resolution and duration. If audio is enabled, the same file includes a synchronized audio track with ambience, sound effects, and any described speech-like audio. You also get basic metadata with the output, such as a generation seed for reproducibility.
Recommended settings
- Choose MP4 when you need maximum playback compatibility across devices and web players.
- Choose MOV when you plan color grading, keying, or compositing. Some players may not support every MOV encoding combination.
- Use 720p when you want cleaner detail and fewer compression artifacts.
- Use a first frame when you need exact framing, wardrobe, or product identity.
Limitations
- Output is limited to short clips, with a maximum duration of 30 seconds per generation.
- Output resolutions are limited to 480p and 720p for this Flash tier.
- When you provide a first frame, the model keeps that image’s aspect ratio. This can override a requested ratio.
- Content filters can block or reject requests, including cases involving public-figure likeness or face/voice similarity signals.
- Results can still vary between runs, even with the same settings.
- Low-quality inputs can degrade results. Blurry images, heavy compression, watermarks, screenshots of UIs, and scanned images can cause flicker, warped anatomy, or drifting textures.
- Fast, precise physics can fail in some shots. Break complex actions into simpler beats when needed.
Safety & compliance
Seedance v2.5 Flash runs with automated safety checks. These checks may block content that matches disallowed categories or resembles protected identities. Don’t use it for impersonation, non-consensual deepfakes, or deceptive media. Only upload content you have rights to use, and avoid prompts that request copyrighted characters, logos, or branded audio. If you enable a watermark, it adds an “AI Generated” disclosure mark to the exported video.
Example prompts
Great starting points for seedance v2.5 flash.
API quick start
Run seedance v2.5 flash with a single API call.
{
"prompt": "The surfer carves down the face of the gl…",
"resolution": "480p",
"duration": 4,
"generateAudio": "false"
}curl -X POST "https://api.wiro.ai/v1/Run/bytedance/seedance-v2-5-flash" \
-H "Content-Type: application/json" \
-H "x-api-key: YOUR_WIRO_API_KEY" \
--data-binary @- <<'JSON'
{
"prompt": "The surfer carves down the face of the gl…",
"resolution": "480p",
"duration": 4,
"generateAudio": "false"
}
JSONOther models you might like
Explore similar video generation models.