How to Make the Alors on Danse AI Video

Make it in 3 steps

1. Pick your cast. On the keys is the calm producer at the mini keyboard; the hype is the friend who points, lunges into the lens and raps. Put your most expressive friend on the hype.
2. Upload two photos on the generator. You get a 16:9 landscape video, framed like the original sketch. Swap the roles with one tap if you change your mind. Pay once, from $9.99 — no account.
3. Download and add the sound. Usually ready in 3–8 minutes; we email you the link. The video is silent on purpose — add “Alors on danse” by Stromae from the app's sound library (step-by-step below).

Photo checklist

• One person per photo — no group shots.
• Face large, sharp and facing the camera, both eyes visible. Waist-up is ideal.
• No sunglasses, masks, heavy filters or beauty effects.
• Bold-coloured clothes look best against the white studio; white and beige tend to disappear.
• Only use photos of people who agreed to be in the video.

Add the sound on TikTok, Instagram or CapCut

1. Open the downloaded MP4 in TikTok or Instagram Reels.
2. Tap Add sound and search “Alors on danse” by Stromae.
3. Drag the clip of the song so the chorus kicks in as the close-up lands (about 1.5 seconds in).
4. Turn on the AI-generated content label (TikTok's rules) and post.

In CapCut: import the video, tap Audio → Sounds, search the song, line the chorus up with the lunge and export.

The Alors on danse AI prompt (do it yourself)

If you want to run a video model yourself, this is the exact prompt our generator sends. It expects three reference images in this order: @Image 1 = the person on the keys, @Image 2 = the hype, @Image 3 = an empty white studio with a small table, a mini keyboard and a laptop. Render 10 seconds with audio off.

A comedy sketch filmed in an all-white studio. @Image 3 is the set only: a seamless white void with no visible walls, corners, ceiling or floor line, only a soft round light glow at the top centre; overexposed, soft, almost shadowless light. A small white table holds a compact black mini MIDI keyboard placed right in front of an open silver laptop; keep the table and those props in shot 1. Exactly two people appear: the KEYS PLAYER is the person in @Image 1 and the HYPE MAN is the person in @Image 2. Keep each one's face, hair, facial hair, glasses, skin tone, build and clothes exactly as in their own photo, unchanged in every shot, with every accessory worn exactly as in their photo. Never swap them, never blend their faces and never add anyone else. Ordinary rectangular video: every corner of the frame shows the bright white studio, evenly lit edge to edge, in every shot. Exactly two shots. Exactly ONE hard cut, at 7.0 s; there is no other cut. Shot 1 (exactly 7.0 s, 0.0-7.0, one continuous take): it opens as a static waist-up medium two-shot at eye level, the table with the mini keyboard and the laptop right at the bottom edge of the frame. On the left the keys player stands at the table playing the mini keyboard with both hands, nodding to the beat; on the right the hype man stands close, grinning, and jabs a finger at the keys player. Then at 1.5 s, in the same take with no cut, the hype man walks right up to the camera and leans in so close that the hype man's face fills the right half of the frame while the keys player stays fully visible in the left third; being this close makes the face look slightly exaggerated. From 2.5 s to 7.0 s the hype man raps and hypes straight into the camera with raised eyebrows and a wide open mouth, head bouncing on every beat, while the keys player, still in focus, keeps playing the keyboard, then shakes both hands in the air to the rhythm. The camera stays almost still with a tiny handheld drift. Shot 2 (exactly 3.0 s, 7.0-10.0): both side by side in the same white void, no table; the shot starts as a medium shot framed from the knees up, both heads and faces fully in frame with a little headroom, then the camera pulls back fast until they are small full-body figures with only a soft shadow under their feet; their faces stay visible for the whole shot; all the while they do the same bouncy jogging dance, knees up, both thumbs pointing back at their own chests, big grins. Bright even high-key light, crisp TV-sketch video look, photorealistic, natural skin texture. Nobody holds a phone or a microphone. No text, no captions, no logos, no brand marks on the laptop.

Common fixes

• The two people swap or blend: make the photos more different and keep “Never swap them” in the prompt.
• A round black fisheye frame appears: keep the “full-frame… no circular fisheye border” wording.
• The lunge comes as a cut: keep shot 1 as “one continuous take”.
• Extra people in the background: say “exactly two people” and use an empty studio image.

Or skip the retries: make yours from two photos, and read where the Alors on danse meme comes from.