Seedance 2.0 AI Video Generator
ByteDance’s video model for people, motion and sound in one pass: text, a photo, reference images or two frames in, a clip with audio out.
Write your idea first. You sign in when you press Generate.
Short answer: Seedance 2.0 is ByteDance’s AI video model, launched on February 12, 2026. It makes a video and its soundtrack together, cuts between several shots inside one clip, and takes text, images, video and audio as input. On Genvima you use it in the browser with no API key: write a prompt or add a photo, pick 5, 10 or 15 seconds at up to 1080p, and the clip arrives with sound and without a watermark, usually in two to five minutes. It starts at 66 credits a clip; the cheaper Fast (40 credits) and Mini (19 credits) versions take the same prompts. It is paid with credits; to try an idea first, the free Basic engine in the same menu gives you one free video and three free images a day.
How to use Seedance 2.0 online
Everything happens on this page. You only sign in with Google when you press Generate.
- 1
Pick how the clip starts
Text/Image to Video for a written idea or one photo, Reference to Video to keep a character or a product from up to 9 images, Frames to Video to set the first and the last frame.
- 2
Describe the shot and the sound
Subject, action, camera, then what you should hear. Seedance 2.0 makes audio, so a line in quotes becomes speech and "rain on the window" becomes rain you can hear.
- 3
Choose version, length and size
Seedance 2.0, Fast or Mini; 5, 10 or 15 seconds; 480p, 720p, 1080p. The Generate button shows the price in credits for exactly that combination.
- 4
Generate and download
The clip renders on our side, so you can close the tab. It is saved to Assets as an MP4 you can download and post.
Seedance 2.0 key features
Official reference clips from ByteDance. Suggested prompts are separate from the clips. Vendor showcases may use settings and inputs beyond those offered here.
Family portraits become a reunion scene across several shots. Try a related prompt
Shot 1: wide shot of a rainy neon street at night, a courier on a bike rides toward the camera. Shot 2: close-up of her face as she brakes hard. Shot 3: the parcel slides across a wet counter into a waiting hand. Sound of rain, a bell, tires on wet asphalt.
Multi-shot clips up to 15 seconds
One prompt can hold several shots. Seedance 2.0 cuts from a wide shot to a close-up to a reaction inside the same clip and keeps the same people and place across the cuts, so a 15 second clip can tell a small story instead of showing one moment.
Hands touch glass, fabric and other surfaces with synchronized sound. Try a related prompt
Medium shot, a barista in a busy morning café slides a cup across the counter and says "Oat flat white for Sam!" Steam hisses from the machine, cups clink, low chatter in the background.
Sound generated with the picture
Audio is not added afterwards. Music, ambience, sound effects and voices come out of the same pass as the video, in stereo, so a door slam lands on the door and footsteps land on the step. Put a line of dialogue in quotation marks and the character says it.
A woman gathers tomatoes in a sunlit courtyard. Excerpt of the finished multi-shot film. Try a related prompt
Use @image1 as the woman and @image2 as the jacket. She walks out of a subway station into morning light, stops, and smiles at the camera. City traffic, a distant train.
References: images, video and audio together
ByteDance lets one request carry up to 9 images, 3 video clips and 3 audio clips. The model can take a face from one image, an outfit from another and the camera move from a clip. On Genvima, Reference to Video takes up to 9 images today; name them in the prompt as @image1, @image2 and so on.
A character steps out of a painting in a fictional cola ad. Try a related prompt
Use @image1 as the host. She holds up a ceramic mug in a bright kitchen, turns it to show the glaze, and says "Handmade, one of one." Soft kitchen ambience.
The same character, shot after shot
Seedance 2.0 keeps a subject’s face, clothes and voice stable across the shots of a clip, and a reference image keeps them stable across clips. That is what makes a short series with one host, one mascot or one product possible.
A pair of skaters perform lifts, spins and jumps. Try a related prompt
Low-angle tracking shot, two figure skaters glide side by side, jump together, spin and land in sync. The camera orbits them once. Blades scraping ice, a crowd gasps then cheers.
Camera and motion that follow direction
Write the camera like a director: dolly in, orbit, handheld, crane up. ByteDance tuned the model for hard motion such as dance, sport and several people moving at once, and it keeps weight and timing believable where older models let limbs slide.
A woman hangs laundry in the sunlight, animated from a starting image. Try a related prompt
Animate the starting photo: a woman hangs a white shirt on a balcony clothesline, then lifts another shirt from a basket. Warm sunlight, a light breeze moving the fabric, soft street sounds.
Natural motion from a starting photo
Start with a photo and describe what should move: a person turning, fabric catching the breeze or hands lifting an object. The image sets the scene, while the prompt guides the action. For a specific destination as well as an opening, choose Frames to Video and add a last frame.
Seedance 2.0 specs and credits on Genvima
| Model | Best for | Resolution | Clip length | Sound | Modes | From |
|---|---|---|---|---|---|---|
| Seedance 2.0 | Best overall | 480p, 720p, 1080p | 5, 10, 15 s | Yes | Text/Image, Reference, Frames | 66 credits |
| Seedance 2.0 Fast | Faster | 480p, 720p | 5, 10, 15 s | Yes | Text/Image, Reference, Frames | 40 credits |
| Seedance 2.0 Mini | Best value | 480p, 720p | 5, 10, 15 s | Yes | Text/Image, Reference, Frames | 19 credits |
| Seedance 2.5 | Top quality | 480p, 720p, 1080p | 5, 10, 15 s | Yes | Text/Image, Reference, Frames | 97 credits |
What you can make with Seedance 2.0
Six things the model is good at, each with a prompt you can paste into the box above and change.
Talking characters and UGC-style ads
A person speaking one or two lines to camera, with matching voice and room sound. Useful for product intros and ad tests where filming a person is the slow part.
Try this promptSelfie-style video, a young man in a parked car holds up a phone charger and says "This thing saved my road trip." Natural daylight, light traffic outside.
Photo to video for social posts
Start from one portrait or product photo and let it move: a turn, a smile, wind, a slow push-in. This is how most photo trends are made.
Try this promptAnimate the photo: she turns her head toward the camera, her hair moves in the wind, the camera pushes in slowly. Soft breeze, birds far away.
Short stories in one clip
Use the multi-shot ability for a 15 second scene with a beginning, a turn and an end, like a mini trailer or the opening of a series.
Try this promptShot 1: a boy finds an old radio in an attic. Shot 2: close-up as he turns the dial. Shot 3: the attic fills with warm light and music. Dust, a crackle, then a song.
Product and e-commerce shots
Reference a product photo and ask for a turntable, a pour, an unboxing or a lifestyle shot. Keep the product description exact so labels and shapes stay right.
Try this promptUse @image1 as the bottle. It stands on wet black stone, water drips down the glass, the camera orbits once in slow motion. Deep bass hit at the end.
Dance and trend videos
Motion-heavy formats suit Seedance well. Our ready-made Tiny Desk AI effect, which puts your photos into a desk concert, runs on the Seedance 2.0 family.
Try this promptFull-body shot, a woman in a red tracksuit dances in an empty parking garage, quick footwork and a spin, the camera tracks at waist height. Upbeat electronic beat, sneakers squeak.
Storyboards and previews
Directors and agencies use it to show a scene before a shoot: camera moves, timing and mood in seconds, at Mini prices while ideas are still changing.
Try this promptCrane shot rising over a desert highway at dawn, a lone vintage car drives toward the horizon, long shadows. Wind, a distant engine.
Seedance 2.0 vs Fast vs Mini vs 2.5
The four Seedance versions on Genvima take the same prompts, so you can draft on one and finish on another. What changes is detail, speed, size and price.
| Seedance 2.0 | Seedance 2.0 Fast | Seedance 2.0 Mini | Seedance 2.5 | |
|---|---|---|---|---|
| Best for | Final clips with people and sound | Quick turnarounds | Drafts, volume and trend videos | Hero shots that need the most detail |
| Resolution | 480p, 720p, 1080p | 480p, 720p | 480p, 720p | 480p, 720p, 1080p |
| Clip length | 5, 10 or 15 s | 5, 10 or 15 s | 5, 10 or 15 s | 5, 10 or 15 s |
| Sound | Yes, generated with the video | Yes, generated with the video | Yes, generated with the video | Yes, generated with the video |
| Ways to start | Text/Image, Reference, Frames | Text/Image, Reference, Frames | Text/Image, Reference, Frames | Text/Image, Reference, Frames |
| Credits, 5 s at the lowest size | 66 | 40 | 19 | 97 |
Seedance 2.0 prompt guide
Seedance 2.0 follows long, specific prompts well. A reliable order is: who, doing what, where, how the camera moves, and what you hear.
- Subject and action first. "A chef flips a pancake" beats a list of adjectives. One main action per shot.
- Camera in plain words. Close-up, wide shot, dolly in, orbit, handheld, top-down. The camera line changes the result more than style words.
- Sound last, and specific. Name two or three sounds. For speech, put the exact words in quotes and keep them short: about two words a second fits the clip.
- Several shots: number them. "Shot 1: … Shot 2: … Shot 3: …" with one line each. Three shots fit well in 10 to 15 seconds.
- References: point at them. "@image1 as the woman, @image2 as the room." Say which image is the person and which is the place.
- Prompt length. The box takes up to 4,000 characters, but around 500 is usually the sweet spot: enough for three shots and their sound.
Seedance 2.0 image to video tips
In Text/Image to Video, a photo you add becomes the first frame, and the prompt only has to say what happens next. Describe the motion, not the picture: the model already sees the picture.
- Use a sharp photo with the subject large in frame. Faces smaller than a thumbnail tend to drift.
- Match the clip’s aspect ratio to the photo: a vertical portrait for 9:16, a landscape for 16:9.
- Ask for small, physical motion first (a turn, a breath, wind). Big actions from a still photo are where results vary most.
What ByteDance announced
ByteDance’s Seed team launched Seedance 2.0 on February 12, 2026, first in its own Dreamina and Doubao apps. The launch post describes a model that reads four kinds of input at once (text, images, audio and video), produces 15 second multi-shot clips with two-channel sound, and adds video extension and targeted editing of characters, actions and storylines.
On Genvima you get 5, 10 or 15 second clips, the three ways to start shown above, and sound on every clip.
Limits worth knowing
ByteDance itself lists what still needs work, and it matches what we see:
- Text inside the picture (signs, labels, captions) is often misspelled. Add text afterwards in an editor.
- Crowds and many characters can blend or swap details. Two or three people hold up best.
- Sound occasionally distorts on loud or layered scenes. Ask for fewer sound sources.
- Real people. ByteDance restricts real faces as subject references. A request with a recognizable real person can be refused; a refused or failed clip gives its credits back.
- Length. 15 seconds is the most one clip can be. Longer videos are several clips joined in an editor.
Seedance 2.0 vs Kling and Veo
People often compare Seedance 2.0 with Kling and Veo 3.1, which are also on Genvima. Seedance 2.0 is the one to pick for people, dialogue and multi-shot clips, and it is the cheapest of the three to draft with thanks to Mini.
Veo 3.1 is very strong at realistic light and makes 8 second clips; choose it for a single cinematic shot. Kling 4.0 is good at controlled camera moves and long, steady motion. Prices for all of them are in the pricing table.
FAQ
What is Seedance 2.0?
Seedance 2.0 is ByteDance’s AI video generation model, launched on February 12, 2026. It generates video and sound together, can cut between several shots in one clip, and accepts text, images, video and audio as input.
Where can I use Seedance 2.0 online?
Right here: pick Seedance 2.0 in the model menu at the top of this page, write a prompt or add a photo and press Generate. You sign in with Google and pay in credits; you do not need an API key or a ByteDance account.
How long can Seedance 2.0 videos be?
Up to 15 seconds per clip. On Genvima you choose 5, 10 or 15 seconds, at 480p, 720p, 1080p.
Does Seedance 2.0 generate audio?
Yes. Music, ambience, sound effects and speech are generated in the same pass as the picture, on Seedance 2.0, Fast and Mini. You can describe the sound in the prompt.
Can Seedance 2.0 make characters talk?
Yes. Put the line in quotation marks in the prompt and the character speaks it with matching mouth movement. Keep it short enough to say in the clip, about two words a second.
How much does Seedance 2.0 cost?
From 66 credits for a 5 second clip at the lowest resolution; a 5 second clip at 1080p is 355 credits. Fast starts at 40 credits and Mini at 19. Credits come with a plan or a credit pack.
Is Seedance 2.0 free?
No. Seedance 2.0 is paid with credits: you sign in with Google, buy a plan or a credit pack, and every clip shows its price on the Generate button before you spend anything. To try an idea first, the free Basic engine in the same menu gives you one free video and three free images a day, with no credit card.
What is the difference between Seedance 2.0 and Seedance 2.0 Mini?
Mini has the same features and takes the same prompts at a much lower price (19 against 66 credits), at up to 720p instead of 1080p. Draft on Mini and finish on 2.0.
Is Seedance 2.5 better than Seedance 2.0?
[Seedance 2.5](/video/seedance-2-5) renders finer skin, fabric and light and costs more (from 97 credits). Seedance 2.0 is the better value for most clips and for many variations.
Can I use Seedance 2.0 videos commercially?
Yes. Clips made with a paid plan or credit pack are yours to use in ads, client work and posts, without a watermark. You are responsible for the people, brands and music you put into the prompt or the references.
What happens if a clip fails?
If a model fails or refuses the request, the credits for that clip go back to your balance automatically.
Keep exploring
- Seedance 2.5ByteDance’s newest video model: finer skin, fabric and light, multi-shot stories and sound in one pass.Open
- Tiny Desk AIYou on lead vocals in the viral desk concert, your friends as the band.Open
- Kling 4.0Kuaishou’s newest video model for long, steady camera moves and fast action that stays smooth.Open
- Veo 3.1Google’s video model: realistic scenes with dialogue, sound effects and ambience made in the same pass.Open
- AI Video MakerEvery way to make a video and all 13 video models.Open
- PricingPlans and credit packs. One balance pays for every video and image model.Open
Try Seedance 2.0 now
Write the idea, pick the model, and see the price before you generate.
Back to the prompt box
Seedance 2.0