Close-ups, comedy, 3D animation, extreme sports and sci-fi scenes from the Seedance 2 family. Click Use template on any clip to load its prompt and input image into the generator.
Seedance 2.0 is ByteDance Seed's video model that generates picture and sound together from text, images, audio and video.
Seedance 2.0 launched in April 2026 as ByteDance Seed's flagship video model. Most video models draw the frames first and leave sound for a separate tool. It is built on a unified multimodal audio-video joint generation architecture, so dialogue, ambience and effects are generated in the same pass as the picture and stay in sync with it — lips move with the words, footsteps land on the beat. On ImagesArt you run it without a ByteDance account or an API key.
Speech, ambience and sound effects come out of the same generation as the video. Write the line in quotes and the character says it with matching mouth movement. Audio is always on and costs no extra credits.
Within one 4–15 second clip the model can cut from a wide shot to a close-up and back, with natural transitions, so the result plays like an edited sequence — and the same face, outfit and lighting carry across every cut.
Upload the first and the last frame and Seedance fills every frame between them, so the clip lands exactly where you planned. Good for reveals, transformations and bridge shots.
Fast action, fabric, water and hands hold together instead of melting. The model keeps body motion and object contact believable even in dance, sports and fight choreography.
Tracking shots, orbits, push-ins, whip pans and handheld moves follow your prompt. Describe the camera the way a director would and Seedance plans the move.
Pick any length from 4 to 15 seconds and six aspect ratios — 16:9, 9:16, 1:1, 4:3, 3:4 and 21:9 — or auto to match your image. Draft cheap at 480p, then rerun the winner at 1080p or 4K.
Choose Text to Video, Image to Video or Start-End Frame. Upload a photo (JPG, PNG or WebP, up to 30 MB) or just write the scene.
Say what moves, how the camera travels and what the mood is. Put dialogue in quotes for lip-synced speech, and name any sound effects you want.
Pick a Seedance model, resolution and length, check the credit cost on the Create button, and generate. Clips usually take 3–12 minutes and land in your history as an MP4.
The model follows direction closely. The more your prompt reads like a shot list, the more control you get.
Label each shot — "Shot 1: wide shot of… Shot 2: close-up of…" — and Seedance cuts between them inside one clip. Want one unbroken shot instead? Add "single continuous take, no cuts".
Tracking shot, dolly zoom, rack focus, crane up, whip pan, handheld, slow push-in, overhead. Name the move and where it starts and ends.
Write the exact line in quotes and say who speaks it — the character delivers it with matching lip movement. Keep each line short enough to fit the clip length.
Name the music, ambience and effects: "rain on a tin roof", "crowd cheering in the distance", "low synth drone". Otherwise Seedance picks sound that suits the scene.
Shot 1: slow push-in on a matte black coffee grinder on a marble counter, morning light through the window. Shot 2: close-up of beans pouring into the hopper, crisp crunching sound. Shot 3: overhead shot of fresh grounds falling into a cup as steam rises, soft acoustic guitar in the background.
Handheld selfie-style shot of a young woman in a busy night market, neon signs behind her. She smiles at the camera and says: "You have to try the dumplings here, trust me." Sizzling woks and chatter in the background.
A cyclist rides through a narrow old-town alley at dawn. The camera tracks beside her, then cranes up over the rooftops as she reaches the harbor and the sun breaks over the water. Single continuous take, no cuts. Seagulls, distant church bells.
All three run on ImagesArt with native audio. Here is how they differ on this site.
| Seedance 2.0 | Veo 3.1 Lite | Kling 3.0 | |
|---|---|---|---|
| Clip length | 4–15 seconds | 4, 6 or 8 seconds | 3–15 seconds |
| Resolution | 480p · 720p · 1080p · 4K | 720p · 1080p · 4K | 720p · 1080p · 4K |
| Aspect ratios | 16:9, 9:16, 1:1, 4:3, 3:4, 21:9 + auto | 16:9, 9:16 + auto | 16:9, 9:16, 1:1 |
| Native audio | Yes, with lip-sync | Yes | Yes |
| Text to video | Yes | Yes | Yes |
| Image to video | Yes | Yes | Yes |
| Start & end frame | Yes | Yes | Yes |
| Reference images | Up to 9 | Up to 3 | Up to 7 |
| Billing | Per second | Per clip, any length | Per second |
| 5 seconds at 720p | 205 credits (Fast: 124) | 15 credits per clip | 90 credits |
| Best for | Dialogue, multi-shot scenes, choreography, widest ratio choice | Cheap cinematic 720p clips at a flat price | Character motion and storytelling |
Every model in the family is billed per second of video. Mini and Fast stop at 720p; the full 2.0 model goes up to 4K.
| Model | 480p | 720p | 1080p | 4K |
|---|---|---|---|---|
| Seedance 2.0 Mini | 3.8 credits / sec19 credits for 5s | 8.2 credits / sec41 credits for 5s | — | — |
| Seedance 2.0 Fast | 11.7 credits / sec59 credits for 5s | 24.8 credits / sec124 credits for 5s | — | — |
| Seedance 2.0 | 19 credits / sec95 credits for 5s | 41 credits / sec205 credits for 5s | 102 credits / sec510 credits for 5s | 208 credits / sec1040 credits for 5s |
Audio is included in every price. Save credits by testing your prompt on Mini or Fast at 480p, then render the keeper on the full model at 1080p or 4K. The exact cost is shown on the Create button before you generate.
Buy a one-time credit pack or subscribe for monthly credits — both unlock higher resolutions. 480p is open to everyone, 720p comes with a Starter plan or any credit pack, and 1080p and 4K with Pro or any credit pack.
언제든 해지 가능 · 결제 기간이 끝날 때까지 이용
크레딧
기능
크레딧
기능
크레딧
기능
크레딧
기능
Run Seedance 2.0 next to Veo 3.1, Kling 3.0, Grok Imagine and Gemini Omni Flash on one page with one credit balance — plus reference mode with up to 9 images.
Open the AI image to video generatorNative audio, lip-sync, six aspect ratios and up to 4K — right in your browser, no software to install.
Compare plans and credit packs