A(caps)I video generation has made it easier to create cinematic dance videos using simple reference photos and a detailed AI video prompt. In this tutorial, you can create a stylish couples dancing video with two reference images, cinematic lighting, premium outfits, smooth camera movements and a detailed seven-scene storyline.
The complete process is beginner-friendly. You only need two clear reference photos, Google Flow and the video prompt provided below. Follow the steps carefully to create your own cinematic vertical AI dance video.
Why This AI Dance Video Is Trending
This concept creates a cinematic dance performance on a dark circular wooden stage. Two performers appear together with consistent facial identities, stylish outfits, dramatic overhead lighting and smooth professional camera movements.
The prompt is divided into seven scenes so that the video has a proper visual flow instead of looking like a random AI animation. It includes the entrance, coordinated dance movements, controlled rotations, a dance dip and a cinematic ending.
Join 3.4
What You Need Before Starting
- A clear photo of the male performer.
- A clear photo of the female performer.
- Access to Google Flow.
- Enough generation credits for your selected settings.
- A stable internet connection.
How to Create Trending Couples Dancing AI Video
Step 1 — Copy the Complete Video Prompt
First, click the Copy Prompt button below to copy the complete video prompt. The prompt contains the character details, outfits, environment, lighting, seven scenes, camera movement and final quality instructions.
📋 Complete Video Prompt
Create an ultra-realistic, premium cinematic 4K/8K video in vertical 9:16 format featuring two people performing an expressive, synchronized contemporary dance together on a dimly illuminated circular wooden stage. The visual style should feel like a high-budget fashion film combined with an elegant live dance performance, featuring dramatic overhead lighting, rich natural colors, deep blacks, subtle atmospheric haze, realistic fabric movement, cinematic depth, and smooth professional camera choreography.
CHARACTER IDENTITY — EXTREME CONSISTENCY:
Use the two uploaded reference photos as the exact identity references for the two performers. The first uploaded image belongs to the male performer and the second belongs to the female performer. Preserve each person's exact facial identity, facial structure, bone structure, eyes, eyebrows, nose, lips, jawline, skin tone, natural proportions, hairstyle characteristics, and recognizable features throughout the entire video. Maintain the same individual identity from Scene 1 through Scene 7 with zero facial drift, identity swapping, morphing, blending, or transformation. Their appearance must remain consistent during walking, dancing, spinning, dipping, lifting, and turning. Preserve the natural body proportions and overall build shown in each person's reference image without exaggeration or alteration.
MALE OUTFIT — DISTINCTIVE LUXURY EDITORIAL LOOK:
The man wears a sophisticated midnight-black tailored double-breasted wool jacket with a subtle satin lapel, layered over a premium charcoal-black silk-cotton shirt with the top buttons naturally open. He wears perfectly tailored black pleated trousers with a refined straight-leg silhouette and elegant matte-black leather loafers. Add a very subtle brushed-metal watch and minimal accessories. The outfit should feel modern, expensive, understated, masculine, and fashion-editorial — no conventional suit appearance, no flashy logos, no costume styling.
FEMALE OUTFIT — ELEGANT MODERN COUTURE LOOK:
The woman wears a sophisticated deep emerald-green satin midi dress layered beneath a flowing long ivory-beige tailored coat with a soft structured silhouette. The dress moves naturally with every dance movement. She wears elegant dark-brown ankle boots with a refined silhouette and minimal jewelry consisting of delicate metallic earrings and a subtle bracelet. Her hair remains long, natural, and freely flowing throughout the performance. The overall styling should look like premium contemporary couture — elegant, cinematic, sophisticated, and visually distinctive.
ENVIRONMENT & LIGHTING:
A circular dark wooden performance stage surrounded by an expansive deep-black background. A single powerful soft-edged spotlight hangs directly above the performers, creating a beautiful pool of light around them while the surrounding environment gradually disappears into darkness. Add extremely subtle atmospheric haze so the light rays have cinematic depth. Preserve rich full-color grading: warm natural skin tones, emerald and ivory clothing highlights, deep blacks, realistic wood tones, soft silver-white highlights, subtle warm shadows, and premium filmic contrast. Never make the image dull, grey, faded, or excessively desaturated.
SCENE 1 — THE ENTRANCE:
Wide cinematic shot. The two performers slowly walk together onto the center of the circular wooden stage while gently holding hands. Their movement is natural, calm, and synchronized. The man's tailored jacket moves subtly as he walks, while the woman's flowing coat follows naturally behind her. Her long hair moves softly with each step. The overhead spotlight gradually reveals them from the darkness. The camera makes a very slow, elegant forward movement, establishing the stage and creating a cinematic atmosphere.
SCENE 2 — EYE CONTACT:
Medium cinematic shot. They stop at the center and turn toward each other. Their hands remain gently clasped. They look naturally at each other, communicating emotion through subtle expressions rather than exaggerated acting. Her hair softly frames her face while the light creates delicate highlights along both faces. Hold the moment briefly with almost imperceptible camera movement and shallow cinematic depth of field.
SCENE 3 — SHARED MOVEMENT:
Wide full-body shot. They gradually raise their joined hands together in a graceful contemporary-dance gesture. Both display subtle genuine smiles. Their movement should feel spontaneous yet perfectly coordinated. The woman's ivory coat and emerald dress respond naturally to the motion, while the man's tailored jacket follows his body movement realistically. The camera gently arcs around them while maintaining both performers fully inside the frame.
SCENE 4 — CONTINUOUS 360° SPIN:
Dynamic full-body cinematic shot. The man carefully and smoothly supports the woman as he lifts her into a controlled contemporary dance spin. He rotates continuously through one complete 360-degree turn while maintaining stable posture and realistic balance. Her feet leave the stage naturally, while her flowing ivory coat and emerald dress flare outward from centrifugal motion. Her long hair sweeps outward and creates elegant curved motion trails without becoming distorted. The camera simultaneously performs a smooth controlled orbit around the pair, capturing the entire rotation without cuts, sudden acceleration, camera shake, or unnatural body deformation. Both performers remain physically consistent and anatomically realistic throughout the complete turn.
SCENE 5 — FREEDOM:
Wide cinematic shot. They return naturally to standing position and extend their arms outward while continuing to hold hands. They lean gently away from each other, creating a beautiful symmetrical contemporary-dance pose. The woman's coat moves softly in the surrounding air, while her hair flows naturally. The man maintains a confident but relaxed posture. The overhead spotlight creates a striking circular pool of light beneath them. The camera slowly pulls backward, emphasizing the isolation of the stage against the black environment.
SCENE 6 — ELEGANT DANCE DIP:
Dramatic wide full-body shot. The man smoothly guides the woman into a controlled classic dance dip, supporting her securely with both arms. She leans backward gracefully while maintaining a natural expressive pose. Her long hair falls naturally toward the wooden floor, while her ivory coat drapes elegantly downward and the emerald fabric catches the spotlight. One leg extends naturally with the boot pointed slightly outward. Hold the pose for a brief cinematic moment before he carefully and smoothly brings her back upright. The movement must remain controlled, believable, graceful, and physically realistic.
SCENE 7 — FINAL ROTATION & EMBRACE:
Final cinematic sequence. The man securely lifts the woman into his arms in a controlled dance movement. He performs one smooth continuous 360-degree rotation while carrying her. Her long hair, flowing coat, and dress move naturally with the rotation, creating elegant cinematic motion. They exchange warm smiles while maintaining clear individual facial identities. As the rotation completes, he gently lowers her back onto the stage and they transition naturally into a close final dance pose. Her hair gradually settles around her shoulders while the fabric slowly returns to its natural position. Finish with a subtle slow cinematic zoom-out as the overhead spotlight softly surrounds them and the rest of the stage disappears into darkness.
CAMERA & MOTION:
Use professional cinematic camera movement throughout: slow dolly-ins, controlled tracking, graceful orbital movement, smooth crane-like motion, subtle push-ins, and elegant slow zoom-outs. Every movement should feel physically believable and fluid. No jerky animation, no sudden camera jumps, no unnatural acceleration, no frozen limbs, no duplicated body parts, no distorted hands, and no abrupt transitions.
IDENTITY & CONTINUITY PRIORITY:
The two performers must remain exactly the same individuals throughout all seven scenes. Never change their faces, facial proportions, skin tones, body proportions, hairstyles, or recognizable characteristics. Maintain consistent clothing, accessories, lighting direction, stage design, and overall visual identity from beginning to end. No face morphing, no identity blending, no character replacement, no outfit changes, and no unexpected additional people.
FINAL VISUAL QUALITY:
Ultra-realistic premium cinematic photography, physically accurate lighting, realistic skin and hair, detailed fabric simulation, natural shadows, realistic wooden stage texture, volumetric spotlight, shallow depth of field where appropriate, high dynamic range, filmic contrast, subtle cinematic grain, rich natural color grading, realistic motion blur, professional dance-film cinematography, 4K/8K detail, vertical 9:16 composition.
IMPORTANT: Exactly 7 scenes only. Preserve the complete story flow from entrance to final dance pose. Do not skip, merge, or replace any scene. No text, no captions, no logos, no watermark, no branding, and no visible written elements anywhere in the video.
✓ Prompt copied successfully!
Step 2 — Open Google Flow
After copying the prompt, click the Create Video button. It will open Google Labs in a new tab. From there, open Google Flow and start your video creation workflow.
Step 3 — Create a New Project
After opening Flow, sign in with your Google account if required and create a new project. The interface may change as the tool is updated, so the exact position or name of buttons can be different.
Step 4 — Upload Your Two Reference Photos
Click the available + or image upload option and add your two clear reference photos. Upload the male photo first and the female photo second, matching the order described in the prompt.
For better consistency, use clear images where the faces are easily visible. Avoid extremely blurry images or photos with heavy filters.
Step 5 — Paste the Video Prompt
Paste the complete prompt that you copied from this article into the appropriate generation field. Make sure the complete seven-scene prompt is included.
Step 6 — Select Ratio and Quality
For YouTube Shorts, Instagram Reels and other vertical platforms, select 9:16 if the option is available. Choose the available quality and generation settings according to your account and available credits.
Step 7 — Generate the Video
Once the reference photos, prompt and settings are ready, start the generation process. Depending on the selected model and reference images, the result can vary, so another generation may be needed if the first result does not match your expectations.
Step 8 — Preview and Download
After the video has finished generating, preview the complete result. Check the faces, hands, clothing, dance movements, camera motion and scene continuity. If the result looks good, use the available download option to save the generated video.
Tips for Better AI Dance Video Results
- Use high-quality and clear reference photos.
- Keep both faces clearly visible.
- Use the same reference photos if generating again.
- Do not remove important sections from the prompt.
- Use 9:16 for Shorts and Reels.
- Preview the complete video before publishing.
- If the result contains unexpected changes, try generating again with better reference images.
Why a Detailed Prompt Is Important
A simple instruction such as “create two people dancing” gives the AI a lot of freedom to decide the characters, environment, camera movement and choreography. This can sometimes produce inconsistent results.
A detailed prompt gives the model a much clearer creative direction. Here, the character identities, outfits, stage, lighting, seven scenes, camera movements and final visual style are described separately.
The seven-scene structure also creates a clear visual progression, making the video feel more like a planned cinematic sequence rather than a random AI-generated clip.
Final Words
Creating a cinematic couples dancing video from two ordinary photographs is now much easier with modern AI video-generation tools. With the right reference images and a carefully structured prompt, you can create a visually engaging dance sequence featuring customized characters, stylish outfits, dramatic lighting, and cinematic camera movements.
The key is not simply generating a video, but giving the AI enough creative direction to understand the characters, environment, movement, camera, lighting, and overall visual style.
Follow the steps above, upload your two clear reference photos, paste the complete prompt, select 9:16, choose a suitable quality based on your available credits, and generate your video.
Frequently Asked Questions
