Seedance 2 prompt guide

Seedance 2 Prompt Guide

Five copyable prompts for Seedance 2.0, one for each way the model is driven on Imgo: text with native audio, a first frame, a first and last frame, reference images and a fast vertical draft. Each one opens in the video generator with the model, aspect ratio and length already set.

Standard, Fast and Mini

The three tiers take the same inputs and all generate native audio. What changes is the resolution ceiling, the clip lengths on offer and the price. Every value below is read from the model registry that each request is validated against.

CapabilitySeedance 2.0Seedance 2 FastSeedance 2 Mini
Resolution480p / 720p / 1080p / 4K480p / 720p480p / 720p
Per-second rate at 720pBaseline (official rate)About 55% of Seedance 2.0 (estimated rate)About 50% of Seedance 2.0 (estimated rate)
Clip length5 / 6 / 8 / 10 / 12 / 15s5 / 6 / 8 / 10 / 12 / 15s4 / 5 / 6 / 7 / 8 / 9 / 10 / 11 / 12 / 13 / 14 / 15s
Aspect ratios16:9, 9:16, 1:1, 4:3, 3:4, 21:916:9, 9:16, 1:1, 4:3, 3:4, 21:916:9, 9:16, 1:1, 4:3, 3:4, 21:9
Native audioYesYesYes
Start / end frameYesYesYes
Reference images444
Reference videoYesYesYes
Reference audioYesYesYes

Input rules that shape the prompt

  • A request runs on frames or on references, never both: reference images cannot be combined with a start or end frame.
  • An end frame needs a start frame. A start frame alone is image-to-video; a start and end frame together is a transition.
  • A reference request takes up to 4 images, up to 3 video clips and up to 3 audio tracks.
  • Native audio is on by default. Write the sound into the prompt, or switch audio off when you will add your own track.
  • In our own test runs, Seedance's content review rejected start frames that show a recognizable face, AI-generated faces included. Frame people from behind, or keep the face out of the first frame.

Credits and failed runs

Seedance 2 is quoted per second of video, and the generator shows the exact credit cost before every run. Native audio carries no separate charge. When a generation fails, its credits go back to your balance automatically.

Imgo pricing →

Copy and adapt

Dialogue scene with sound effects

A 10-second 16:9 scene in [setting]. [character], [description], [action]. Medium shot, then a slow push-in. Dialogue in [language]: the character says, "[line]". Sound effects: [sound effect] exactly as [action beat] happens. Ambience: [ambience]. Lips move in sync with the line. No music, no subtitles, no on-screen text.

Mode
Text to video · native audio · 16:9 · 10s
Use it when
The clip has to speak: a line, a sound effect and the room tone written into the shot instead of added in the edit.
Inputs
Text only. Keep the line short enough to say comfortably inside the clip.
Check
Check lip sync, the spoken words, and that the effect lands on the action the prompt ties it to.

Product image to video

Use the uploaded image as the first frame. Keep [product], its label, logo, colors and proportions exactly as shown. Over 8 seconds the camera [camera move] while [ambient motion] moves gently in the background; the product itself stays rigid. One continuous shot: no cuts, no morphing, no new objects, no text changes. Audio: [ambience] and a soft [product sound]; no music, no voiceover. End on a stable two-second hold.

Mode
Image to video · first frame · 9:16 · 8s
Use it when
A product photo is already approved and needs to move without being redrawn.
Inputs
One product image in the target aspect ratio, with no recognizable face in frame; it becomes frame one.
Check
Compare the opening frame with your image, then watch the label, logo and edges for drift as the camera moves.

First frame to last frame

Create an 8-second transition from the first-frame image to the last-frame image as one continuous shot. Keep [subject] identical throughout. Motion: [transition action], while the camera [camera move]. Timing: hold the opening for one second, move steadily, and settle on the final frame for the last second. Audio: [transition sound] building under the motion and landing on a soft [final sound] as the last frame settles; no music.

Mode
Image to video · first and last frame · 16:9 · 8s
Use it when
The start and end states are fixed — a before and after, a reveal, a transformation — and only the path between them is open.
Inputs
Two images with the same aspect ratio, framing and lighting; the closer they match, the cleaner the path.
Check
Confirm the clip really lands on the last frame and that the subject keeps its identity halfway through.

One subject across a new scene

Use the reference images as the exact [subject]: keep its shape, colors, markings and proportions unchanged in every shot. A 10-second 16:9 sequence in [setting] with [lighting]. Open on a slow push-in, then [subject action], and end on a centered hero frame with clean negative space on the right. Audio: a [music style] music bed and [sound effect] in sync with the action; no voiceover. No on-screen text and no extra subjects.

Mode
Reference images · up to 4 · 16:9 · 10s
Use it when
A product or character has to stay exactly right while the set, the camera and the sound are invented around it.
Inputs
Two to four images of the same subject from different angles. No start or end frame in this request.
Check
Compare shape, color and markings against the references in every shot, not only the first.

Vertical hook draft

A 5-second 9:16 opening shot for a short-form ad about [product]. In the first second, [pattern interrupt] fills the frame at close range, then the camera snaps back to reveal [product] in [setting]. Handheld energy, natural light, real textures. Audio: a sharp [hook sound] on the first frame, then [ambience]; no music, no voiceover. No on-screen text, no logos other than the product's own.

Mode
Text to video · Seedance 2 Fast · 9:16 · 5s
Use it when
You are testing openings for a short-form ad and want several cheap takes before paying for a final render.
Inputs
Text only. Run it several times, keep the strongest take, then render that one in Seedance 2.0.
Check
Judge the first second on a phone with the sound on: does the motion stop the scroll, and does the sound hit with it?

Tested on Imgo

Each clip below is the unedited output of a prompt from Imgo's prompt packs, generated on Imgo with the model named on the card. Open the prompt to see exactly what produced it.

Macro product reveal

Seedance 2.0 · 9:16

Pull continuously from material detail into a full hero composition.

See the prompt that made this →

Shot 4 · Product hero hold

Seedance 2.0 · 9:16

Push in, then hold on the label inside the safe zone.

See the prompt that made this →

Shot 1 · Pattern-interrupt hook

Seedance 2 Fast · 9:16

Product in the very first frame; the first 1.5 seconds decide the view.

See the prompt that made this →

Practical tips

  • Decide between frames and references before you write: the two cannot share a request, and the prompt reads differently for each.
  • Write sound like a shot list: who speaks, which effects, which ambience, and whether there is music. Asking for no music is an instruction too.
  • Pick a clip length the tier offers. Seedance 2.0 and Fast render 5, 6, 8, 10, 12 or 15 seconds; Mini renders any whole second from 4 to 15.
  • Draft in Fast or Mini at 720p and keep 1080p and 4K for the take worth finishing — only Seedance 2.0 renders above 720p.
  • Say what must not change — label, logo, colors, proportions — and end with the exclusions: no cuts, no new objects, no on-screen text.

Frequently asked questions

What is the best prompt structure for Seedance 2?
Subject and action first, then the camera, then setting and light, then the audio, then what must not change or appear. In a reference request, open by saying what the reference images are for.
Can I use these prompts on Imgo right now?
Yes, when the status note at the top of this page shows the model as live — it reads Imgo's model catalog. Each prompt has a button that opens the video generator with the prompt, model, aspect ratio and length filled in.
What inputs does Seedance 2 accept?
Text alone; a start frame; a start and end frame; or references — up to 4 images, up to 3 video clips and up to 3 audio tracks. Frames and reference images cannot be mixed in one request. Clips run 5 to 15 seconds, or 4 to 15 on Mini.
What is the difference between Seedance 2.0, Fast and Mini?
The inputs and native audio are the same. Seedance 2.0 renders up to 4K. Fast and Mini stop at 720p and cost roughly half as much per second at 720p — about 55% for Fast and about 50% for Mini, both estimated rates. Mini also offers every length from 4 to 15 seconds.
Do I lose credits when a Seedance 2 generation fails?
No. Credits are reserved when the run starts and returned to your balance automatically if the generation fails.

Reviewed on October 14, 2026 against Imgo's model registry. The credit quote in the generator is the price source of truth.