How-to2026-11-16•6 min read

How to Make the Jingle Bell Brazil Video With Your Photo

Add your photos, pick solo, duo or trio, check the free preview, and generate the 15-second video with the original track — seats, photo rules and fixes.

Step 1 — Add your photos

You need one photo per seat you want to fill. The rules that actually decide whether the result looks like you:

  • Front-facing, chest-up, face sharp and fully visible — the rebuild works from the face it can see.
  • Solo or Duo? Go waist-up and show your top. Those two versions keep the outfit from your photo — you dance in your own clothes — so a shot that shows what you are wearing beats a tight head shot. Trio is the opposite: it dresses you in the original Santa-hat outfit, so chest-up is all it needs.
  • Nothing covering the face: no sunglasses, no mask, no hand in front of it, no hair across the eyes.
  • Bright, even light. One warning specific to this clip: the trio room is lit by blue LED strips, and that light darkens and blues faces — a bright, well-lit photo holds up far better than a dim one.
  • No heavy beauty filters. A filter that reshapes the face fights the rebuild, and the result lands between the two.
  • One person per photo. A group shot leaves the generator guessing which face belongs in the seat.
  • Adults only, and only with their consent — the same rule the generator itself enforces.

The seats, in detail

Every version of the dance is a fixed line of dancers, and each photo replaces the dancer in one seat. Seats are named from the viewer’s side of the screen: left, center and right, exactly as you see them in the clip. Solo has one seat — the single dancer in the living room. Duo has two, left and right, in the store Christmas aisle. Trio has all three in the original blue LED room.

Filling every seat is optional, and that is the point of the system. Fill only the center seat of the trio and you dance between the two original masked dancers. Fill the left seat of the duo and your friend keeps the original partner on the right. Each upload box belongs to one seat — the generator numbers them in order, left, center, right — and you can remove and re-upload any single seat without touching the others. Switching versions does not clear your photos either: each version keeps its own set, so you can set up the trio, peek at the duo, and come back to find the trio untouched.

Step 2 — Preview it free, then generate

Before you spend anything, press "See your free preview". You get a still of yourself in the actual scene — free once you sign in, one preview a day (a preview spends real image-model calls, which is why it is tied to an account). The still is not a mock-up: it is built from the same reference photos and the same frame of the base clip that the video render uses, so the face, the costume and the seat you see are the ones in the finished video. If a photo is too dark or a seat assignment is wrong, this is where you find out — swap the photo and preview again.

When the preview looks right, pick your version and generate. Solo is the living-room take — one dancer, one photo. Duo fills one or both seats in the store Christmas aisle. Trio is the original blue-room line. There is no prompt to write: the edit prompt is pre-loaded and tuned beat by beat — stillness until about 3 seconds, the cross-and-chest-tap, the hand-to-mouth move, the main dance, and the freeze at the end.

The render edits the fixed base clip rather than generating a new scene, so the room, the side dancers, the choreography and the timing stay exactly as in the original — and the original track is kept, timed to the dance. A render usually takes 3–5 minutes on the same page.

When a render goes wrong

Most bad results trace back to the photo, not the dance. The failures worth knowing:

  • A face looks partly unchanged in a few frames — the pipeline detects this and runs a repair pass on its own; if it persists, the source photo was too dark or too filtered. A brighter photo fixes it more reliably than re-rolling the same one.
  • The wrong person is in a seat — photos apply in seat order, left, center, right, one per upload box. Check which box holds which photo before generating; the free preview shows the assignment before you pay.
  • The result looks dark or very blue — that is the original room’s light, not a defect. A brighter source photo keeps more facial detail under it.
  • The render failed outright — its credits come back automatically, every time. Retry with a different photo; very dark, blurry or heavily filtered photos are the usual cause.
  • It is taking longer than five minutes — busy periods queue up. Leave the tab open; if you close it, reopen the page and the render picks up where it left off.

Step 3 — Download and post

You get a 15-second 9:16 MP4, watermark-free, with no caption sticker baked in — add your own caption in TikTok, Reels or Shorts. It is vertical on purpose: the line of three fills the frame, and a wide crop would cut the side dancers out.

Two posting tips. Label the video as AI-generated — the platforms ask for it and the trend’s audience expects it. And the render carries the original track for watching and sharing as-is; if you post publicly and want the trend’s reach, attaching the sound from TikTok’s own sound page is the route most versions take.

Read next

Where the clip came from, the remix behind it, and how this generator compares with the others that appeared the same week.