SaWaDiKa dance AI video generator
Step into the SaWaDiKa solo routine with one photo. The 15-second vertical dance reference and prompt are ready to use: add your character, choose a resolution and generate your remake.
One person, clear front-facing photo.
What is the SaWaDiKa dance?
"SaWaDiKa" is a single by LISA, released on September 3, 2026 through Lloud and RCA Records as the lead single of her EP Press Play. The title is the Thai greeting "sawatdi kha" — hello — and the song is a tribute to Thailand. The music video, shot in Bangkok, includes a rooftop dance with her crew.
Dancers posted the chorus moves as soon as it dropped, and AI versions starring friends and pets followed. This template keeps the 15-second solo routine and puts the person from your photo in LISA's place.
How this template works
- 01Add your character photo
Upload one clear photo as @Image1. Choose a single person with a visible face and the hairstyle and outfit you want in the remake.
- 02Use the original dance reference
The bundled 15-second solo clip attaches automatically as @Video1. The prompt asks to keep its motion, pose, expression timing, camera movement, framing, lighting and background.
- 03Pick a resolution and generate
Choose 768p for 240 credits or 2K for 390 credits. MiniMax Hailuo 3, the prompt, 15-second duration and 9:16 framing are preset.
Your face, the same dance brief
- 01One photo to start
No need to film yourself dancing or prepare a motion upload. The solo reference and character-replacement prompt are included.
- 02Identity from your image
The prompt uses your photo for facial features, face shape, hairstyle, skin tone and outfit, with natural blending into the scene's lighting and color grading.
- 03A complete vertical routine
Start from the full 15-second, 9:16 reference. Review the face, fast gestures and scene continuity before sharing your result.
Prepare your references
One character photo
Upload a JPEG, PNG or WebP photo with one clearly visible person. Keep the face unobstructed and choose the outfit you want the performer to wear.
The bundled SaWaDiKa routine
The 15-second vertical reference is included automatically. It stays locked with the prompt and 9:16 output so you can focus on the character photo.
What to check in the result
- Compare facial features, face shape, hairstyle, skin tone and outfit at the start, middle and end.
- Replay fast hand and body movements to check for flicker, distorted limbs and unstable face edges.
- Check dance timing, expression changes, camera movement, framing and background against the reference.
Character identity, motion, camera and background can change during generation. Exact face replacement, unchanged backgrounds and lip sync are not guaranteed. Begin with one short test before making a larger batch.
FAQs
Do I need a dance video or a prompt?
No. Upload one character photo. The SaWaDiKa motion reference, replacement prompt, model, duration and framing are already configured.
Which parts come from my photo?
Your photo defines the new performer's face, hairstyle, skin tone and outfit. The video guides the dance, pose and expression timing, camera, lighting and background. AI results can vary; check identity and motion throughout the clip.
What kind of photo works best?
Use a sharp, well-lit photo of one person, facing the camera, with the intended clothing visible. A full-body or upper-body standing photo gives more context than a tightly cropped face.
Will the original song audio be included?
This template is for a visual remake; the original song audio is not included. Add your chosen audio in your editor after reviewing the generated video.
Related briefs and guides
How to replace a person in a video with AI
Prepare a short reference video and a character photo, write a clear reference prompt, review the quote and inspect a character-led video remake.
Choose a character photo for AI video remakes
A practical checklist for framing, pose, lighting, clothing and choosing a reference photo that fits a short video remake.