Multimodal input stackDetails
Seedance 2.0 accepts text, image, audio, and video references.
- Text instructions + multimodal references
- Up to 9 image references
- Up to 3 video references
- Up to 3 audio references
bytedance model
Seedance 2.0
Seedance 2.0 is ByteDance's next-gen AI video model focused on cinematic motion, multi-shot continuity, and native audio generated in sync with visuals. This page is a pre-launch overview: specs, best use cases, and prompt templates so you can plan workflows before release.
Best for Action sequences with more believable physics and interaction, Multi-shot ads with consistent product and scene continuity, and Music-led visuals with synced ambience and SFX cues.
Text→VideoImage→Video1080p15s16:9 / 9:16 / 1:1Audio
Join waitlist · Official date TBASee Seedance 1.5 Pro
Pay-as-you-go on MaxVideoAI · Price shown before you generate (at launch)
- PriceTBD (confirmed at launch)
- Duration15s
- Format16:9 / 9:16 / 1:1
Use cases
Action sequences with more believable physics and interactionMulti-shot ads with consistent product and scene continuityMusic-led visuals with synced ambience and SFX cuesStoryboard-to-video using text and multimodal referencesDirector-style camera moves, transitions, and framing controlPre-visualization before final edit (concept to shot list to rough cut)
What makes Seedance 2.0 different
- Director-level camera language (Optimized for explicit shot direction: camera movement, framing, and transition verbs)
- Multi-shot continuity in a single render (Designed for short timelines where multiple beats can be stitched into one coherent output)
- Native audio generation (visual and sound together) (Official materials highlight audio generated alongside video for stronger sync)
- Multimodal references at scale (Mix text with image, video, and audio references to lock style, pacing, and continuity)
Specs (pre-launch)
The limits that shape your renders.
Price / secondTBD (confirmed at launch)
Text-to-VideoSupported
Image-to-VideoSupported
Video-to-VideoSupported
First/Last frameNot specified
Reference image / style referenceSupported
Reference videoSupported
Max resolution1080p
Max duration15s
Aspect ratios16:9 / 9:16 / 1:1
FPS options24
Output formatMP4
Audio outputSupported
Native audio generationSupported
Lip syncSupported
Camera / motion controlsAdvanced
WatermarkNot specified
Release dateComing soon (official date TBA; availability to be confirmed at launch)
Output style and structureDetails
Model messaging emphasizes cinematic control and multi-shot outputs.
- Up to 1080p
- Up to 15s per generation
- 24 FPS (as announced)
- Natural transitions across multiple shots
- Native audio-video joint generation
- Audio-video sync mentioned in official materials
- Dual-channel (stereo) audio support highlighted by ByteDance
Seedance 2.0 examples
Explore official references and prompt patterns while runtime remains locked pre-launch. We will publish real MaxVideoAI renders here once the model is live. View all Seedance examples ->
Prompt Lab — Seedance 2.0
Official Seedance 2.0 pageSeedance 2.0 tends to reward shot timing, camera verbs, and audio cues. Keep dialogue short, pin SFX to visible actions, and use references to lock continuity.
Tip: duration + aspect ratio are set in the UI — your prompt controls subject, action, camera, lighting, style, and sound.
Quick concept prompt
Fast ideation with one cinematic beat.
Quick = iterate concept and mood.
Template (copy/paste)
[Subject] in [setting]. Camera: [move + lens feel]. Lighting: [style]. Action: [one clear beat]. Audio: [ambience + one SFX cue]. (<=15s, choose aspect ratio in UI.) Example: Handheld UGC unboxing at a kitchen table. Slow push-in, natural daylight. She peels the seal, smiles, turns the bottle to camera. Room tone + packaging crinkle + soft click when cap opens.
Example
Handheld smartphone UGC clip of a woman unboxing a new skincare bottle at a kitchen table. She peels the seal, smiles, and turns the bottle toward camera. Soft window daylight, natural colors, subtle room tone + packaging crinkle.
Demo prompt — Seedance 2.0
Explore official references and prompt patterns while runtime remains locked pre-launch. We will publish real MaxVideoAI renders here once the model is live.
Tips and boundaries
What works best
- Strong cinematic camera direction when beats are explicit.
- Improved continuity with multimodal references.
- Native audio sync in short multi-shot outputs.
- High utility for ads, action beats, and storyboard prototyping.
Common problems → fast fixes
- Cuts feel abrupt -> add transition verbs and timestamps (match cut at 5s, whip pan into Shot 2).
- Continuity drifts -> add anchors (wardrobe, prop, location) and reuse references.
- Audio mismatch -> shorten dialogue, pin SFX to visible action, keep one ambience bed.
- Physics looks off -> simplify simultaneous actions per beat and reduce fast interactions.
Hard limits to keep in mind
- Pre-launch visibility does not imply runtime access.
- Launch-day endpoint exposure determines available advanced operations.
- Commercial details and pricing finalize at launch.
- Use only officially documented capabilities in production planning.
Compare Seedance 2.0 vs other AI video models
Not sure if Seedance 2.0 is the best fit for your shot? These side-by-side comparisons break down the tradeoffs — price per second, resolution, audio, speed, and motion style — so you can pick the right engine fast.
Each page includes real outputs and practical best-use cases.
openai
Seedance 2.0 vs OpenAI Sora 2
Create rich AI-generated videos from text or image prompts using Sora 2. Native voice-over, ambient effects, and motion sync via MaxVideoAI.
Compare Seedance 2.0 vs OpenAI Sora 2 →pika
Seedance 2.0 vs Pika 2.2 Text & Image to Video
Generate stylized AI video from prompts or animate uploaded stills using Pika 2.2. Perfect for short-form loops without audio via MaxVideoAI.
Compare Seedance 2.0 vs Pika 2.2 Text & Image to Video →bytedance
Seedance 2.0 vs Seedance 1.5 Pro
Generate Seedance 1.5 Pro clips with cinematic motion, camera lock, and native audio. Supports text-to-video or image-to-video up to 12s.
Compare Seedance 2.0 vs Seedance 1.5 Pro →Safety & people / likeness
- Don’t generate real people or public figures (celebrities, politicians, etc.).
- No minors, sexual content, hateful content, or graphic violence.
- Don’t use someone’s likeness without consent.
- Some prompts and reference images may be blocked — generic characters and scenes are fine.
FAQ
What is Seedance 2.0?
Seedance 2.0 is ByteDance's AI video model focused on cinematic motion, multi-shot continuity, and native audio generation.
How long can outputs be?
Launch materials highlight outputs up to 15 seconds per generation.
Does it support audio?
Yes. Native audio is highlighted in official messaging, with audio generated alongside video for better sync.
When will it be available on MaxVideoAI?
This page is pre-launch. Availability and pricing will be confirmed at launch (official date TBA).
Can I generate before launch?
No. The model page can be indexed for discovery, but runtime remains locked until release.