Intro video · How to Make an AI Dance Video: Prompt Formula, Timed Beats and a Photo That Dances (1 min 17 sec)
AI Dance Video Prompts to make people, characters and photos dance
Every example on this page is a real video we generated, shown with the exact prompt, model, settings and credits. You’ll find a 10-second solo routine in timed beats, a 15-second crew shot list, 3D, anime and penguin dancers, ballet and salsa, a photo that starts dancing, and two AI Motion Control videos that copy the solo routine onto a new person and a cartoon character. For dancing cats and dogs, see our cat and dog video prompts.
The court routine was made with Seedance 2.5 at 720p and the crew routine with Seedance 2.5 at 480p. The 5-second examples used Wan 3.0, and the last two used AI Motion Control in Bandicam AI Studio, with the court routine as the dance video. All people, animals and characters are AI-generated, and no real songs, dances or artists are named. The intro video at the top was edited from these examples and a Bandicam screen recording.
Written and tested by Bandicam AI Team · Last updated · Every example video was made in Bandicam AI Studio
Fill in these six parts. The examples come from the sunset court routine below. For a 5-second clip, one line of moves is enough; for 10 seconds or more, split the moves into timed beats.
1
Dancer
Age, hair and outfit. For a character or animal, say what it is and how it looks.
A woman in her late 20s with a high curly ponytail, an oversized mustard-yellow hoodie
2
Moves
Two or three concrete moves, or one per timed beat. Vague words like “freeze” or “cool moves” are easy to misread.
grooves on the beat with bouncing shoulders and step-touch footwork
3
Place and light
Where it happens and the light. A floor with room to move helps.
An outdoor basketball court at golden hour
4
Framing and camera
Vertical and full-body, so the feet stay in frame. A static camera makes the moves easiest to follow.
Vertical full-body dance video, static camera, her head and feet stay in frame the whole time
5
Music by mood
Genre, tempo and mood instead of a song title, plus “instrumental” and “no vocals.”
Upbeat synth-pop instrumental at about 120 BPM with handclaps, no vocals, no lyrics
6
What to leave out
Text, logos and extra people you don’t want on screen.
No text, no subtitles, no logos, no watermark, no other people
Full body, static camera
A dance reads only when the feet are in frame. Write “vertical full-body” and “static camera,” and for longer clips add “her head and feet stay in frame the whole time.”
Name concrete moves
“Step-touch footwork,” “one controlled turn” and “a smooth body roll” are easier to follow than “freeze” or “cool moves.” Two or three moves fit in 5 seconds.
Split 10 seconds or more into beats
Write timed ranges like “0-3s:” with one or two moves each. The court routine and the crew shot list on this page came out in the written order.
Describe the music, don’t name it
Write the genre, tempo and mood, such as “upbeat synth-pop instrumental at about 120 BPM,” and add “no vocals, no lyrics.”
From a short prompt to a full prompt
The short prompt most people start with and the full prompt behind one of the examples, side by side with the video each one made.
Short promptFull prompt
Writing only “a woman dancing” let the model choose everything: a woman in a white dress turning in a bright studio, with the camera drifting closer until her feet leave the frame. The full prompt behind the “Full prompt” video names the dancer, one move (a slow pirouette on pointe), the stage and light, a full-body frame and the music, and her feet stay in frame. The colored phrases match the parts of the formula (the color key is above the examples).
We took one example from this page, changed a single part and generated it again. Everything else in the prompt, the model and the settings stayed the same. The highlighted words are what changed. When a result is not what you want, change one thing at a time like this.
Only the dancers’ age changed in the prompt. The white shirt, red dress, string lights and tracking camera carried over, and the couple still turns with the skirt swirling; the square and the dress came out a little different.
Changed prompt · 367 characters
Vertical full-body shot: a couple in their 70s, a man in a white linen shirt and a woman in a flowing red dress, dance salsa in a warm-lit town square at night, with quick turns and her skirt swirling. The camera slowly tracks around them, string lights overhead. Lively Latin percussion and trumpet instrumental. No vocals, no lyrics, no dialogue. No text, no logos.
Model
Wan 3.0
Aspect ratio
9:16
Length
5 sec
Resolution
480p
Credits
48
November 21, 2026 · Wan 3.0 · first result, published as generated
in a colorful toy room→on the surface of the moon, Earth glowing in the sky
Only the place changed in the prompt. The moon version still bounces, swings its arms and spins, with Earth overhead and dust underfoot, though the robot’s details differ slightly from the toy-room version.
Changed prompt · 394 characters
Vertical full-body shot: a chubby 3D-animated orange robot with round glowing blue eyes dances on the surface of the moon, Earth glowing in the sky, bouncing its knees, swinging its arms and spinning once on the beat. Soft 3D cartoon render, bright playful lighting, static camera. Bouncy electronic dance instrumental and little toy beeps. No vocals, no lyrics, no dialogue. No text, no logos.
Model
Wan 3.0
Aspect ratio
9:16
Length
5 sec
Resolution
480p
Credits
48
November 21, 2026 · Wan 3.0 · first result, published as generated
AI dance video prompts: 10 examples
Click a thumbnail to play the example video. Every prompt is shown exactly as it was used to make the video, so you can copy it or send it straight to Bandicam AI Studio.
The colors show which part of the prompt formula above each phrase belongs to.
1Dancer
2Moves
3Place and light
4Framing and camera
5Music by mood
6What to leave out
CutTiming and cuts
01 · Static camera · four timed beats · instrumental
Sunset court routine: a 10-second solo in four beats
The camera is static and the prompt says “her head and feet stay in frame the whole time,” so the whole routine stays readable. The 10 seconds are split into four beats with one or two concrete moves each, and they came out in that order: grooving, a turn and arm poses, side steps, and a hand-on-hip finish. This clip is also the dance video for the two AI Motion Control examples below.
Prompt · 986 characters
[Subject] A woman in her late 20s with a high curly ponytail, an oversized mustard-yellow hoodie, black cargo pants and white high-top sneakers.
[Place] An outdoor basketball court at golden hour, a chain-link fence and a glowing city skyline behind her.
[Style] Vertical full-body dance video, static camera, her head and feet stay in frame the whole time, warm sunset light, crisp motion, realistic.
[Must exclude] No text, no subtitles, no logos, no watermark, no other people.
[Dance beats]
0-3s: she grooves on the beat with bouncing shoulders and step-touch footwork.
3-6s: she makes one controlled turn, then snaps her arms into three sharp poses on the handclaps.
6-8s: she takes two quick side steps and does a smooth body roll, still facing the camera.
8-10s: she stops in an upright final pose with one hand on her hip, holds it and smiles.
[Sound] Upbeat synth-pop instrumental at about 120 BPM with handclaps, no vocals, no lyrics. Sneaker squeaks on the court on the beat.
Model
Seedance 2.5
Aspect ratio
9:16
Length
10 sec
Resolution
720p
Sound
On
Credits
740
November 21, 2026 · Seedance 2.5 · first result, published as generated · loose curls instead of a high ponytail
02 · Five shots · three dancers · one action per shot
Parking-garage crew: a 15-second shot list
Each 3-second range gets one shot and one action: a wide shot, a sneaker close-up, a top-down shot, a popping solo and a slow-motion jump. All five came out in the written order. Small details can drift: the woman’s hair came out as a platinum buzz cut instead of a silver bob, and the two men behind her keep moving instead of holding still.
Prompt · 1107 characters
[Subject] Three adult dancers in matching black-and-white streetwear: a tall man with short twisted hair, a woman with a silver bob, and a man with a red bandana.
[Place] An empty underground parking garage at night, wet concrete floor, cool fluorescent lights and one red neon strip on the wall.
[Style] Vertical dance video, high contrast, sharp motion, realistic.
[Must exclude] No text, no subtitles, no logos, no watermark, no other people.
[Shot list]
0-3s: wide full-body shot, the three dancers stand in a triangle and hit the first beat together.
3-6s: low close-up of their sneakers doing quick footwork in sync, small splashes on the wet floor.
6-9s: overhead top-down shot, the dancers circle around one another like a turning wheel.
9-12s: medium shot, the woman in front does sharp, robotic popping moves with her arms and chest while the two men hold a still pose behind her.
12-15s: slow-motion wide shot, all three jump and land in a final pose as the neon flickers.
[Sound] Hard-hitting hip-hop instrumental with deep bass, no vocals, no rap, no lyrics. Sneaker squeaks echo in the garage.
Model
Seedance 2.5
Aspect ratio
9:16
Length
15 sec
Resolution
480p
Sound
On
Credits
494
November 21, 2026 · Seedance 2.5 · first result, published as generated · platinum buzz cut instead of a silver bob; the men behind her keep moving in the popping shot
03 · 3D character · three moves · static camera
A 3D robot dancing in a toy room
“3D-animated” and “soft 3D cartoon render” set the look without naming any studio. The robot bounces, swings its arms and spins once, the three moves listed in the prompt.
Prompt · 364 characters
Vertical full-body shot: a chubby 3D-animated orange robot with round glowing blue eyes dances in a colorful toy room, bouncing its knees, swinging its arms and spinning once on the beat. Soft 3D cartoon render, bright playful lighting, static camera. Bouncy electronic dance instrumental and little toy beeps. No vocals, no lyrics, no dialogue. No text, no logos.
Model
Wan 3.0
Aspect ratio
9:16
Length
5 sec
Resolution
480p
Sound
On
Credits
48
November 21, 2026 · Wan 3.0 · first result, published as generated
04 · Anime style · clearly adult character · room to move
An anime office worker dancing after hours
“Adult” and “in his 30s” keep the anime character clearly grown up, and “a clear space between the desks” leaves room to move. In 5 seconds the result is a light side-to-side groove with swinging arms while papers flutter around him; the stubble in the prompt barely shows.
Prompt · 426 characters
Vertical full-body anime-style shot: an adult office worker in his 30s with light stubble and a loosened tie, alone in an empty office at night, dances in a clear space between the desks, stepping side to side and swinging his arms as loose papers flutter around him. Cel-shaded 2D anime look, blue monitor glow, static camera. Upbeat funk instrumental and soft footsteps. No vocals, no lyrics, no dialogue. No text, no logos.
Model
Wan 3.0
Aspect ratio
9:16
Length
5 sec
Resolution
480p
Sound
On
Credits
48
November 21, 2026 · Wan 3.0 · first result, published as generated
05 · Animals · matching props · hard floor for tap sounds
Tap-dancing penguins on ice
Matching bow ties make the three penguins read as a group, and “a smooth, glassy sheet of ice” gives the taps a hard surface. They lift their feet and spread their wings roughly together.
Prompt · 343 characters
Vertical full-body shot: three penguins in tiny black bow ties tap-dance in a line on a smooth, glassy sheet of ice, flapping their wings and stamping their feet in sync. Static camera, bright blue sky, sparkling snow around the ice. Cheerful swing-band instrumental with crisp tap sounds. No vocals, no lyrics, no dialogue. No text, no logos.
Model
Wan 3.0
Aspect ratio
9:16
Length
5 sec
Resolution
480p
Sound
On
Credits
48
November 21, 2026 · Wan 3.0 · first result, published as generated
06 · Ballet · one move · slow motion
Slow ballet turns on an empty stage
The prompt asks for one slow pirouette on pointe in slow motion instead of fast turns, and the legs stay clean. The result makes two turns as the camera slowly pushes in, and several stage lights appear rather than a single spotlight.
Prompt · 367 characters
Vertical full-body shot: a ballerina in her 30s in a white tutu performs one slow pirouette on pointe alone on an empty theater stage under a single spotlight, dust floating in the light. Slow motion, slow push-in, the dark, empty theater around her. Soft solo piano instrumental and the creak of the stage floor. No vocals, no lyrics, no dialogue. No text, no logos.
Model
Wan 3.0
Aspect ratio
9:16
Length
5 sec
Resolution
480p
Sound
On
Credits
48
November 21, 2026 · Wan 3.0 · first result, published as generated · two turns instead of one; several stage lights instead of one spotlight
07 · Partner dance · two different outfits · tracking camera
A salsa couple in a town square at night
Two clearly different outfits, a white linen shirt and a red dress, keep the partners easy to tell apart during turns. The camera tracks around them as the skirt swirls. Where partners hold hands, check the hands closely in your own results.
Prompt · 370 characters
Vertical full-body shot: two dancers in their 20s, a man in a white linen shirt and a woman in a flowing red dress, dance salsa in a warm-lit town square at night, with quick turns and her skirt swirling. The camera slowly tracks around them, string lights overhead. Lively Latin percussion and trumpet instrumental. No vocals, no lyrics, no dialogue. No text, no logos.
Model
Wan 3.0
Aspect ratio
9:16
Length
5 sec
Resolution
480p
Sound
On
Credits
48
November 21, 2026 · Wan 3.0 · first result, published as generated
Starting image (AI-generated example photo, Nano Banana 2.1, 8 credits)
08 · Image to video · describe only the moves
Make a photo dance
Made from a photo. The prompt doesn’t describe the man again; it says only how he moves and asks for his head and feet to stay in frame. He bounces in place, takes small side steps and points at the camera, and his face stays the same. The starting image is an AI-generated photo.
Prompt · 275 characters
The man in the photo starts dancing in place: he bounces to the beat, takes a few side steps, then points at the camera with a big grin. The camera stays still and his head and feet stay in frame. Upbeat pop instrumental. No vocals, no lyrics, no dialogue. No text, no logos.
Model
Wan 3.0
Aspect ratio
9:16
Length
5 sec
Resolution
480p
Sound
On
Credits
48
November 21, 2026 · Wan 3.0 · first result, published as generated
Starting image (AI-generated example photo, Nano Banana 2.1, 8 credits) + the court routine as the dance video
09 · AI Motion Control · photo + dance video
AI Motion Control: the court routine in a kitchen
AI Motion Control takes two inputs: the photo shown here and the 10-second court routine (the first example on this page). The text prompt only asks to keep her face, clothes and kitchen. She follows the routine beat for beat, from the arm poses to the hand-on-hip finish. To try it, upload a photo and a dance video in AI Motion Control.
Prompt · 104 characters
Keep her face, clothes and the kitchen background from the photo. Follow the dance moves from the video.
Model
AI Motion Control
Aspect ratio
9:16
Length
10 sec
Resolution
720p
Sound
On
Credits
380
November 21, 2026 · AI Motion Control · first result, published as generated · 380 credits (quoted 382)
Character image (AI-generated, Nano Banana 2.1, 8 credits) + the court routine as the dance video
10 · AI Motion Control · character image + dance video
AI Motion Control: the same routine on a 3D character
The same court routine on a 3D character image. The character was drawn with human-like proportions and long arms and legs so the human moves fit. The hoodie, sneakers and plain background stay the same and the tail swings along with the steps, but its big round eyes turn half-closed.
Prompt · 112 characters
Keep the 3D cartoon look of the dinosaur and the plain studio background. Follow the dance moves from the video.
Model
AI Motion Control
Aspect ratio
9:16
Length
10 sec
Resolution
720p
Sound
On
Credits
380
November 21, 2026 · AI Motion Control · first result, published as generated · 380 credits (quoted 382)
The values we chose when we made each example video. Credits show the charge for one generation of that video, or a quoted price where a note below the table says so.
Problems we ran into while making the examples on this page, with fixes from our tests and the model makers' official guides. The last column gives the source.
Problem
Fix
Source
The feet get cut off
Write “vertical full-body” and “static camera,” and for longer clips add “her head and feet stay in frame the whole time,” as in the court routine.
Examples on this page
Moves come out of order or get skipped
Split the clip into timed ranges with no gaps and one or two moves each. The court routine and the crew shot list came out in the written order.
Seedance official guide · examples on this page
Arms or legs blur during spins
Ask for one spin, or slow it down with “slow motion.” Avoid stacking a spin on top of a moving camera.
Examples on this page
Singing or talking shows up
Write “instrumental” for the music and end with “No vocals, no lyrics, no dialogue.”
Wan official guide · examples on this page
The face in my photo changes while dancing
Don’t describe the person again; say only how they move, as in “Make a photo dance.” Keep the moves in place and facing the camera.
Runway, Wan and Seedance official guides
AI Motion Control copies the moves poorly
Use a dance video with one person, a static camera and the full body in frame, like the court routine. A character with arms and legs shaped like a person’s takes the moves best.
Examples on this page
How to make an AI dance video
01
Pick a routine
Choose a text example and click “Use this prompt” to open it in Bandicam AI Studio. Change the dancer, outfit and place, and keep the framing and music lines.
02
Match the model and length
Use Seedance 2.5 for 10- or 15-second routines in timed beats and Wan 3.0 for 5-second clips. Choose 9:16 for Shorts, Reels and TikTok.
03
Make a photo dance
Click “Use my photo” on the photo example, upload your picture and describe only the moves. To copy one exact routine onto a photo or character, upload the picture and a dance video in AI Motion Control.
Try it yourself
The prompt from the “Sunset court routine: a 10-second solo in four beats” example is already filled in. Edit it or write your own, then click “Generate video” to continue in Bandicam AI Studio.
AI dance video prompts: frequently asked questions
How do I make an AI dance video?
Describe the dancer, two or three concrete moves, the place, a vertical full-body shot with a static camera, and the music by mood. For 10 seconds or more, split the moves into timed beats, one range per line, like “0-3s: she grooves on the beat.” Paste the prompt into Bandicam AI Studio, choose a model and length, and generate.
Can I make a photo dance with AI?
Yes, in two ways. With image to video, upload the photo and describe the moves, as in “Make a photo dance” on this page. With AI Motion Control, upload the photo and a dance video, and the person in the photo follows the moves, as in the kitchen example. Use photos of yourself or of people who have agreed.
What is AI Motion Control?
A Bandicam AI Studio tool that transfers body movement from a reference video to a character image. The two examples here used our 10-second court routine as the video and cost 380 credits each. Check the cost shown in the tool before you generate.
How do I keep the moves on the beat?
Give a tempo such as “about 120 BPM,” ask for a clear beat like handclaps, and put the moves into short timed ranges. AI video does not lock moves to the beat exactly, so generate a few times and keep the best take.
Can I use a real song or a trending dance?
The prompts on this page describe the music by mood and the moves in plain words instead of naming a song, artist or choreography. To use a song you have the rights to, replace the audio in a video editor after you download the video. To copy a dance with AI Motion Control, use a dance video you filmed yourself or have permission to use.
Why do hands or legs look strange during fast moves?
Spins and fast footwork are the hardest part. Ask for one spin or slow it down with “slow motion,” keep the full body in frame and the camera still, and generate again if the limbs blur.
What kinds of AI dance videos can I make?
A dancer or crew from a text prompt, an animated or 3D character, an animal, a photo that starts dancing, and one routine copied onto someone else with AI Motion Control. This page has an example of each. For dancing cats and dogs, see our cat and dog video prompts.
Which models were used, and how many credits does it cost?
Court routine: Seedance 2.5, 720p, 10 seconds, 740 credits. Crew routine: Seedance 2.5, 480p, 15 seconds, 494 credits. 5-second examples: Wan 3.0, 480p, 48 credits each; the dancing photo also needed an 8-credit AI photo. AI Motion Control: 380 credits each for a 10-second result, plus an 8-credit starting image. These are November 2026 prices; check the cost shown in AI Studio before you generate.
A New Year's Eve countdown and ball drop, fireworks that spell 2027, a midnight toast and resolution ideas. Vertical videos for Reels, TikTok and Stories.
A rooftop band video, a 15-second synthwave shot list, a singing shot, a singer photo lip-synced to a song, and lofi, lyric video and visualizer backgrounds.