Viaduto Classic
One photoYou, in your own clothes, owning the packed baile under the viaduct.
Best source: One adult, front-facing, waist-up or full-body, clear light.
Viral trend · Baile charme · Updated October 20, 2026
Charme Dance AI video from one photo
Upload a clear photo, choose a charme dance direction, and generate a short vertical clip of the baile charme step in a packed Rio night crowd. yito turns the camera, crowd and motion into one prompt, and Studio shows the live credit quote before you generate.
Updated October 20, 2026
What is the charme dance AI trend?
Charme dance AI is a short vertical AI video in which a person dances the charme step inside a packed night-time baile charme, usually the Viaduto de Madureira party in Rio de Janeiro. The clip is filmed like a handheld phone in the crowd: a glance back, a side-to-side groove, a friend sliding in, the signature charme step and a turn while the crowd raises their hands. The trend spread on TikTok in October 2026 through Seu Zé de Madureira, an AI-made grandpa, and then through remakes with famous faces. yito makes the clip from one photo you have the right to use; it does not ship the CapCut template or the original audio.
Capability
What yito can and cannot do
This page is a starting point, not a one-tap CapCut template. Here is the exact boundary so you can judge the result before you spend credits.
One photo in, one short clip out
Upload one photo with one adult, or two adults for a duo direction. The image-to-video workflow animates that photo; you do not need to find a reference video.
Duration and sound depend on the model
Models return different clip lengths, and only some generate audio. Studio shows the duration, resolution, available audio options and binding credit quote before you submit.
The crowd is synthetic
Background dancers and faces are generated by the model. They are not real people and they change on every render. Review faces, hands and crowd edges before posting.
Six directions
Pick a charme dance direction
Each card opens the same image-to-video workflow with a different prompt. The prompt is prefilled, so you only add your photo and confirm the quote.
Preto e Branco
One photoThe same night, in grainy black-and-white film with hard flash falloff.
Best source: One adult with a clear silhouette and visible outfit.
80s Action Look
One photoFaded denim, a white tank top and a wide belt, like the remakes that went viral.
Best source: One adult; the outfit is generated, so simple clothing works best.
Vovô no Baile
One photoYou at 75, still the best dancer at the baile.
Best source: One adult with the full face visible; age changes can drift.
Casal no Charme
Two people in one photoA couple in the crowd, dancing the charme step together.
Best source: Both adults already in one photo, both faces visible.
With Your Friend
Two people in one photoTwo adults, two dancers: your friend joins you at your shoulder.
Best source: Both adults already in one photo, both faces visible.
The 12-second shape
The beat sheet behind the trend
The viral clip is built on five beats. Use this sheet to judge a render or to rewrite the prompt for a different model.
| Time | Beat | What happens | Prompt cue |
|---|---|---|---|
| 0.0–1.5s | The look back | The phone is at your shoulder in the crowd; you glance back as the beat drops. | glance back, handheld phone at shoulder height |
| 1.5–3.0s | The groove | The camera eases back to waist height; your forearms swing and your shoulders bounce. | side-to-side step, shoulders bouncing, forearms at the waist |
| 3.0–6.0s | The friend | A dancer slides in at your shoulder and you trade a look mid-step. In the duo directions, that is the second person in your photo. | friend enters at the shoulder, shared look |
| 6.0–9.0s | The charme step | You lean forward with soft knees and slide your hands past each other at waist height. | soft knees, hands sliding past each other |
| 9.0–12.0s | The turn | The camera drifts to your side, you turn back to the front and keep the step going as the crowd raises their hands. | turn to camera, crowd hands up |
The sheet describes the motion direction, not a guaranteed duration. The model you pick may return a shorter clip; Studio shows the exact duration before you submit.
How it works
From one photo to a charme clip
- 01
Choose a photo
Use one adult, front-facing, with the face, hair and outfit visible. For a duo direction, use a photo that already contains both adults. No sunglasses, no masks, no celebrities, all adults 18 or older.
- 02
Open a direction and read the quote
Pick one of the six cards. The prompt is prefilled in Studio; choose a model, duration and resolution, review the audio options the model supports, then read the binding credit quote.
- 03
Generate and review
Check identity, hands, feet and the crowd before you post. If the face drifts, try a second render or a different model. The crowd and background faces will not match between renders.
Source photos
What makes a charme dance photo work
The model copies what it can see. A sharp, evenly lit photo is the difference between a recognizable clip and a stranger in the crowd.
Works well
- One adult, or two adults for a duo direction
- Face clearly visible and looking toward the camera
- Bright, even light; daylight or a soft key light
- Waist-up or full-body, with the outfit visible
- A simple background with no bystanders
Avoid
- Group photos with people you do not want in the clip
- Sunglasses, masks or hats that hide the face
- Blurry, dark or heavily filtered selfies
- Extreme angles or a tiny face in the distance
- Anyone under 18, celebrities or public figures
- Anyone who has not agreed to appear
Straight talk before you upload
This page is an independent guide. It does not ship the original CapCut template, the party audio or a guaranteed 12-second result.
- Not affiliated with CapCut, TikTok, the Baile Charme de Madureira, the Viaduto de Madureira, Seu Zé de Madureira or any artist.
- No original party audio is included. Add sound in your editor and check the rights before you publish.
- Clip length and audio depend on the model and settings you choose; 12 seconds is the trend’s shape, not a promise.
- Two-person directions need both adults in the same source photo. yito accepts one reference image per generation.
- Background faces are synthetic and can look wrong. Fast motion can blur hands and feet.
- Celebrity photos, public figures and anyone under 18 are not allowed.
- A generated clip can still drift from the source identity. Review it before posting.
Copy the recipe
The base prompt, without the template
Prefer another editor? Copy the base instruction and adapt the crowd, outfit and camera. The six cards above are variations of this prompt.
Base prompt
Animate the uploaded photo into a 9:16 vertical handheld-phone video at a packed night-time baile charme under the Viaduto de Madureira in Rio de Janeiro. Preserve the main person’s face, facial proportions, skin tone, hair, body proportions and clothing. The person dances the charme step: side-to-side steps with soft knees, shoulders bouncing on the beat, forearms swinging in front of the hips, a glance back at the camera, then a turn to the front while the crowd raises their hands. Warm floodlights, haze, a dense crowd of dancers filming with phones, slight handheld shake, shallow depth of field, realistic skin texture. Keep exactly one main person. No text, no logos, no watermark, no extra limbs, no distorted hands.
See the quote before you generate
yito uses credits. The exact cost depends on the model, duration and resolution you choose. Open a direction, review the live quote in Studio, and only then generate. Credit packs and plans are on the pricing page.
See credit pricingCharme dance AI questions
What is the charme dance AI trend?
It is a short vertical AI video of a person dancing the charme step inside a packed night-time baile charme, usually the Viaduto de Madureira party in Rio de Janeiro. The clip looks like a handheld phone recording from the crowd and spread on TikTok in October 2026.
Can I make a charme dance video from one photo?
Yes, if the photo shows one adult clearly, or two adults for a duo direction. yito animates the photo with an image-to-video model; you do not need a reference video.
Is this the CapCut charme dance template?
No. yito does not ship the CapCut template or the original audio. The six directions are prompt recipes for yito’s own image-to-video workflow.
How long is the clip, and does it have sound?
It depends on the model and settings. Studio shows the duration and whether audio generation is available before you submit. The trend’s 12-second shape is a direction, not a guarantee.
Can I put two people in the video?
Yes, with the Casal no Charme and With Your Friend directions. Both adults must already be in the source photo; yito accepts one reference image per generation.
Can I use a celebrity photo?
No. Use only photos of yourself or people who have given permission. Celebrities, public figures and anyone under 18 are not allowed.
Why does my face look different?
Fast crowd motion and a busy background make identity drift more likely. Use a sharp, front-facing photo, review the render, and try a second render or a different model if the face changes.
How much does a charme dance video cost?
The quote depends on the model, duration and resolution. Studio shows the binding credit quote before you submit; the pricing page explains credit packs and plans.
Is my photo private?
New generations are private by default in yito. Review the privacy policy for retention, deletion and support access details.
Can I post the clip on TikTok, Reels or Shorts?
Yes, if you have the rights to the source photo and the result. Check the music rights separately; yito does not include the original party audio.
Is yito affiliated with the baile charme or CapCut?
No. yito is an independent AI creative workspace and is not affiliated with, endorsed by or sponsored by the Baile Charme de Madureira, the Viaduto de Madureira, CapCut, TikTok, Seu Zé de Madureira or any artist.
Keep exploring
Independent tool
yito is an independent AI creative workspace. This page is an editorial guide to a trend, not an official partner page. Names are used only to describe the trend. Preview images are AI-generated illustrations, not customer results.