Make a video from a prompt, a still image, or reference files. You can try Seedance 2.0 on Topview without a credit card.
Seedance 2.0 is ByteDance Seed's video model. It generates the picture and the soundtrack together from a text instruction, and that instruction can also use images, audio, and video as references.
Those references can guide composition, motion, camera movement, effects, and sound. ByteDance's public launch notes describe up to 9 images, 3 video clips, and 3 audio clips with the text, 15 reference files in total, and they include extending or editing a clip in the same model. On Topview, Seedance 2.0 outputs at up to 4K. You start from a prompt, a still image, or reference files. More prompt examples are in the Seedance 2.0 prompts. The examples below are clips made in that workflow.
These showcase categories are reordered around what paying users actually make most often: film-style scenes, fashion and beauty content, music and dance, ads, viral social concepts, and explainers. Each clip below keeps the prompt used for that example. More starting prompts are in the 50+ Seedance 2.0 prompt templates.
Cinematic narratives with professional pacing. Explore modern tactical operations, mythological encounters, sci-fi thrillers, and surrealist social satire.
Tactical Breach and High-Speed Escape
A modern tactical team breaches a corridor with an explosion. Transition to a man in a suit escaping via elevator to a garage, driving off in a silver sedan while the team exits in a pickup.
The Spartan's Celestial Encounter
A Spartan-style warrior stands on a cliff edge. A massive glowing humanoid silhouette descends from the sky, reaching out a divine finger toward the warrior's raised sword.
The Morgue Anomaly: Sci-Fi Thriller
A female doctor uncovers a corpse in a morgue to find a terrifying rift lined with teeth on the neck. The sequence ends with a sudden, unsettling twitch of the corpse's fingers.
Urban Warfare: Sunset Vigil
A soldier survives a building explosion in a narrow alley, moves through thick smoke into ruins, and ends on a rooftop keeping watch against a dramatic city sunset.
These examples show how Seedance 2.0 handles motion, light, transformations, faces, product presentation, and camera movement. The prompt under each clip is the prompt used for that example.
Fast action stays coherent - clean timing, believable contact, and fewer weird body glitches.
"CG style, cinematic 3D animation. Female soldier, short blonde hair, full tactical gear, combat boots, holding assault rifle. Snowy industrial ruins background, debris, cold atmosphere. High-octane action sequence character sliding on snow while shooting, performing martial arts kicks on masked enemies, sprinting. Dynamic camera angles, motion blur, Unreal Engine 5 render, hyper-realistic lighting, 8k resolution."
Get film-like framing and lighting with more consistent color and depth - less "flat AI look."
"A 1920s-style modern woman with a black bob and pearls walks down a tiled basement staircase at night. The camera tracks backward just below her chin, capturing her expression under soft, directional overhead lighting. The scene features a dreamlike film noir aesthetic with blue-green tones, mist, and balanced exposure, ensuring her face is contoured but visible."
Magical outfit swaps and style transitions that maintain perfect body consistency.
"Using the content structure of this [video] as a reference, generate a costume-change effect video for this character [image1]"
Subtle expressions read better on camera - more believable emotion, fewer "dead face" moments.
"A two-part commercial video sequence. Shot 1: An elegant woman stands in a bright, modern indoor room, extending her hand to hold a glass serum bottle close to the camera lens while talking and explaining the product details. Focus on the bottle texture and her engaging expression. Shot 2: The scene cuts to the interior of a luxury car with leather seats. The same woman sits in the driver's seat, holding a small bottle of essential oil up towards the camera, continuing her presentation with natural daylight streaming through the car windows."
Generate high-converting product showcases with realistic human movements.
"The character from Image 1 picks up the bag from the ground, rotating it to present a 360-degree view. They then hold it forward to showcase the handle details and decorative charm. Finally, they stand up, sling the bag over their shoulder, and pose to demonstrate the upper body fit while introducing the product."
Stunning aerial views of historical landmarks with smooth camera paths and high-fidelity textures.
"Photorealistic 15s FPV drone shot, wide-angle 16mm lens, continuous \"one-take\" style with high-tension movement and motion blur. The camera dives from high-altitude backlit clouds, seamlessly transitioning to a fast spiral close to the Eiffel Tower, weaving through its steel beams. It then cuts to a low-altitude rush through the arches of the Roman Colosseum, pulling up vertically into a bright sun flare."
Create a complete clip in one pass so visuals and sound land on the same beat with less cleanup in post. This is one of Seedance 2.0's biggest workflow advantages over prompt-only competitors.
One-pass | Synced beats | Less post-fix
Type the scene once and generate a structured sequence with cuts, angles, and pacing for up to 15 seconds in one pass. It is built for short narrative arcs, not just one isolated shot.
Up to 15s | Fewer stitches | Smoother arcs
Stage scenes where multiple characters speak and react naturally, making conversations feel more watchable and less synthetic. This helps Seedance 2.0 work for ads, explainers, and story-first creator content.
Natural speech | Reactions | Watchable
Switch quickly from cinematic drama to anime or ad-clean visuals without rebuilding your prompting workflow from scratch. Teams can test more concepts inside one production system.
Cinematic cuts | Guided pacing | Cohesive flow
Use Text to Video when you only have a description. Use Image to Video when a still should move. Use Multi Reference when images, video, or audio should guide the clip.
Say what should happen and what should stay the same. If you add files, say what each file is for.
Start the render, then download the file when it is ready. In one test, a text-to-video clip at 9:16, 1080p, and 5 seconds took about 120 seconds. Other settings and queue conditions can take a different amount of time.
These are the core Seedance 2.0 specs most relevant to creators evaluating output quality, control depth, and production fit.
Dual Branch Diffusion Transformer
About 30% faster than the previous version
Combined across all modalities
Flexible output options for faster drafts or higher-detail delivery
Extendable via multi-shot chaining
Cinema-standard output
6 supported formats
Style, character, and scene references
Motion transfer and camera reference
Beat sync, lip-sync, and atmosphere
Detailed scene, pacing, and direction control
8+ languages, auto-generated
Seedance 1.5 Pro is an earlier Seedance release. Seedance 2.0 is the model on this page. Seedance 2.5 is the newer model, and it is the version Topview currently features first. Stay on this page for Seedance 2.0. Open Seedance 2.5 for the newer release, or read the Seedance 2.5 vs Seedance 2.0 comparison for a version-by-version page. A lighter Seedance 2.0 option is Seedance 2.0 Mini.
Compared with Seedance 1.5 Pro, Seedance 2.0 expands the reference inputs, raises the output ceiling, and brings native audio and editing into the same workflow.
| Capability | Seedance 1.5 Pro | Seedance 2.0 | Improvement |
|---|---|---|---|
| Max Resolution | 480p | 4K | Up to 4K |
| Generation Speed | Baseline | 30% faster | Fewer wait times |
| Max Duration | 5-10s | 4-15s | Longer story arcs |
| Image Inputs | Up to 3 | Up to 9 | 3x more references |
| Video Inputs | Not supported | Up to 3 clips | New capability |
| Audio Inputs | Not supported | Up to 3 files | New capability |
| Total Mixed Inputs | Max 3 | Max 15 files | 5x input capacity |
| Native Audio | Not supported | Dialogue, SFX, lip-sync | Less post-production |
| Video Editing | Not supported | Replace, add, remove, extend | New editing layer |
| Aspect Ratios | 3 formats | 6 formats | 2x more flexibility |
| Architecture | DiT | Dual Branch Diffusion Transformer | Next-gen stack |
| Multi-shot Storytelling | Limited | Full multi-camera storytelling | Narrative quality |
| Character and Physics Control | Basic | Advanced face, clothing, style, and physics lock | Production-grade consistency |
Choosing the right AI video model matters. This comparison focuses on the features that most affect creators, marketers, and production teams in 2026: duration, native resolution, multimodal inputs, audio, editing, speed, and overall workflow control.
| Feature | Seedance 2.0 | Sora 2 | Kling 3.0 | Veo 3.1 |
|---|---|---|---|---|
| Developer | ByteDance | OpenAI | Kuaishou | |
| Max Duration | 15s | 12s | 10s | 8s |
| Max Resolution | 4K | 1080p | 1080p | 1080p |
| Native Audio | Dialogue + SFX + lip-sync | Generated audio | Generated audio | Generated audio |
| Image Inputs | Up to 9 | 1 | 1-2 | 1-2 |
| Video Reference | Up to 3 clips | No | No | 1-2 clips |
| Audio Reference | Up to 3 files | No | No | No |
| Multi-Shot Sequences | Yes | Yes | No | Yes |
| Video Editing | Yes | No | No | No |
| Multi-Language Lip-sync | 8+ languages | Limited | Limited | Limited |
| Approx. Cost (10s 1080p) | ~$0.60 | ~$1.00 | ~$0.50 | ~$2.50 |
| Generation Speed | ~2 min | ~3 min | ~1 min | ~2.5 min |
| API Available | Full | Limited | Full | Full |
| Best For | Multimodal control and storytelling | Physics realism | Speed and value | Cinematic polish |
Seedance 2.0 is the option in this comparison with image, video, and audio references together with editing. That combination fits creators and marketers who want those controls in one workflow. For models beyond this table, .
The clips above are different jobs. Seedance 2.0 fits when you already know the scene: a short film beat, a product story, a performance, a campaign cut, a social hook, or a result driven by a file you already have.
Use Seedance 2.0 when the prompt needs framing, camera language, pacing, and more than one angle in the same generation. If you want to plan the shots before you write that prompt, the AI storyboard generator is a separate planning tool.
Seedance 2.0 can carry a model shot, a routine, a jewelry detail, or a product turn. When the job starts from a product image or a product URL and follows an ecommerce template, use the AI product video generator.
Seedance 2.0 can generate the performance visual. If you still need the track, create it with the , then add an audio file when the Seedance generator gives you an audio reference.
One scene often has to become a hook, a launch cut, and a placement variant. Describe those versions in the prompt. Topview's cover campaign types beyond this model page.
Trend hooks, pet clips, skits, and POV shots can be prompted here as short scenes. The is a separate tool for UGC, avatars, and URL-to-video workflows.
Use Image to Video or Multi Reference when you already have a still, a clip, or audio. The is the broader Topview video tool when you are not staying on Seedance 2.0.
Start from a prompt or a file you already have. Generate cinema-grade, physics-accurate video from text or images directly in your browser with Seedance 2.0.
Topview API
Bring cinematic AI video generation into your application with Topview API. Access Seedance 2.0 for text-to-video, image-to-video, and reference-guided workflows with transparent pricing and developer-ready documentation.
15% below listed official price
$0.1285/ second
$0.1512 / second — Listed official price
720p · No reference video · 16:9 · 24fps
Explore Seedance 2.0 APIDocumentationReference rate for the configuration shown. Pricing varies by settings and usage. API pricing is separate from Topview's web generator credits.
