How to Use Seedance 2.0: Step-by-Step Guide with Prompt Tips

Seedance 2.0 is ByteDance's multimodal AI video model. It turns a prompt, images, video clips or audio into a cinematic clip of up to 15 seconds, with sound generated in the same pass. This guide walks through each way to use it, which settings to pick, and how to write prompts that get the shot you have in mind.

You can follow along on SeeVid's AI Video Generator, which runs Seedance 2.0 in the browser — no installation or API key needed.

Before You Start: Pick the Right Mode

Seedance 2.0 has three modes. Choose the one that matches what you already have:

Step 1: Text-to-Video

1. Open the AI Video Generator and choose the Text to Video tab.
2. Select Seedance 2.0 in the model picker.
3. Write your prompt (see the formula below).
4. Choose your settings:
   • Aspect ratio: 16:9 for YouTube, 9:16 for TikTok and Reels, 1:1 for feeds, or Auto.
   • Duration: 4 to 15 seconds. Start with 5 seconds while you test.
   • Quality: 480p for drafts, 720p for most social posts, 1080p or 4K for final delivery.
   • Generate audio: on by default; turn it off if you will add your own music.
5. Click Generate Video. The cost in credits is shown on the button before you start, and most clips are ready in a few minutes.

Step 2: Image-to-Video

Image-to-video keeps your composition, subject and style, and adds motion.

1. Choose the Image to Video tab and select Seedance 2.0.
2. Upload one image to use as the first frame. Upload a second image to set the last frame too — the model creates the motion between them.
3. Describe the motion, not the picture: "The camera slowly pushes in as she turns toward the window and smiles; curtains drift in the breeze."
4. Generate.

Tip: the image already defines how things look, so spend your prompt on movement, camera and sound.

Step 3: Reference-to-Video (the Real Power of Seedance 2.0)

Reference-to-video is what sets Seedance 2.0 apart. Choose the Reference to Video tab and select Seedance 2.0. You can combine:

• Up to 9 images: characters, products, outfits, locations, style frames
• Up to 3 video clips: a camera move, a dance, an editing rhythm to copy
• Up to 3 audio clips: music or a voice to follow

Each upload gets a tag — @image1, @image2, @video1, @audio1 — and you use those tags in the prompt to say what each reference is for. Click a tag under an upload to insert it.

Example:

@image1 walks through the neon market from @image2, wearing the jacket from @image3. Copy the handheld tracking camera from @video1. Cut to a close-up as she laughs. Background music: @audio1.

Be explicit about the role of each reference ("the face from", "the camera move from", "the music from"). Vague prompts like "use these images" leave the model guessing.

Note: reference videos are billed on top of the output, because the model processes them too. The cost shown on the Generate button already includes them.

The Seedance 2.0 Prompt Formula

Strong Seedance prompts cover six things, roughly in this order:

1. Subject — who or what, with one or two defining details
2. Action — what happens, in sequence
3. Scene — where and when, lighting and weather
4. Camera — shot size and movement (tracking shot, dolly in, orbit, handheld, aerial)
5. Style — cinematic, documentary, anime, product commercial, film stock
6. Sound — dialogue in quotes, sound effects, music mood

Example:

A young chef in a white apron flips vegetables in a flaming wok at a night street stall. Sparks rise as steam fills the frame. Medium shot, slow dolly in to a close-up of the wok. Warm tungsten light, shallow depth of field, cinematic. Sizzling oil, crowd chatter, the chef says "One more minute!"

Multi-Shot Prompts

Seedance 2.0 can cut between shots inside one clip. Describe them in order:

Shot 1: wide aerial of a lighthouse at dawn. Shot 2: close-up of the keeper's hands lighting a lantern. Shot 3: the beam sweeps across the dark sea. Calm orchestral score.

Dialogue

Put spoken lines in quotes and say who speaks: The old man looks up and says "You came back." Seedance generates the voice with matching lip movement.

Settings Cheat Sheet

Common Mistakes (and Fixes)

• The character changes between shots. Upload a clear, front-facing reference image and tag it in every shot description.
• The motion is too static. Name the action and the camera move; "a woman in a park" gives the model nothing to animate.
• Too many ideas in 5 seconds. One clear action per shot works better. Increase the duration or split into several clips.
• The result is close but not quite. Keep the prompt, change one detail at a time, and generate again — each run is a new take.

FAQ

Is Seedance 2.0 free to use? New SeeVid users receive free credits after verifying their email, which is enough to try Seedance 2.0. After that, generations use credits from a pack or subscription. See the Seedance 2.0 pricing guide.

How long does Seedance 2.0 take to generate? Usually a few minutes. Longer clips, higher resolutions and reference videos take longer.

Can I use Seedance 2.0 videos commercially? You can use your generations in your own projects. Make sure you have the rights to everything you upload as a reference, and get permission from anyone whose likeness appears.

What is the difference between Seedance 2.0, 2.0 Fast and 2.0 Mini? Fast and Mini are cheaper, quicker versions for drafts and simple shots. Standard Seedance 2.0 gives the best quality and the most control.

Start Creating

Pick a mode, use the prompt formula, and start with a short 720p test. Once the shot looks right, raise the duration or resolution for your final version.

Open Seedance 2.0 on SeeVid and generate your first clip.