Is the flamingo float video real or AI?

Both versions exist. Ilia Prudnikov (cain.aii on Instagram) labelled his Singapore clip as AI content on 29 October 2026. Younes Zarou has not said how the Dubai clip was made, and thousands of comments under it ask. Yours is rendered from your photo: the drop, the float and the hand come from a reference clip, and nobody goes near a real rooftop edge.

How many photos do I need?

One, of one adult facing the camera. This template has a single seat — there is no second person on the float, and a second photo is not accepted.

How long is the video and what format is it?

11 seconds, vertical 9:16, an MP4 at 720p (1080p is an AI upscale of that master), no watermark. It plays on TikTok, Reels and Shorts as it comes.

Does it come with the sound?

Yes — the “hey bro, bro, bro” and the “thank you” are mixed under the picture, which is what makes the clip land. The sound belongs to whoever recorded it; this site grants no rights to it, so treat it as personal posting rather than commercial use.

Can two of us be on the float?

Not on this template. The head is about 25 pixels wide at the start of the clip and under 10 by the end; two faces would be roughly 12 pixels each and neither would read as the person who uploaded. For two people use the Hop Hop Hop AI duo or the 58th Floor two-seat rap.

Can I put my dog or cat on the float?

Not yet. The seat is a human rider in this release, and a pet needs its own prompt branch because the whole body replaces the person, not just the face.

Can I pick a different city?

No. The clip is cut from the Dubai reel, so the skyline is part of the plate. Rendering New York or Singapore would mean re-drawing the towers over the same footage, which is exactly where those versions break.

The face came out looking like the man in the reel. What now?

That is the most common miss on this trend, because the face is small enough for the model to fall back on the reference. Use a closer, sharper, front-facing photo with no sunglasses and no hat, and run it again. A failed render refunds its credits automatically.

How much does one video cost?

5 credits at 480p, 7 at 720p, 12 at 1080p. The $6.99 pack holds 7 credits, so one pack is one 720p video; the $14.99 pack is 16 credits.

Flamingo Float AI · the infinity-pool trend

Put yourself in the flamingo float video

One photo, eleven seconds, a rooftop pool 300 metres up

Upload one photo and get an 11-second vertical clip of the flamingo float reel: you lie back on the giant pink float at the open edge of a Dubai rooftop infinity pool, phone in hand, until a hand reaches in and pushes you back from the drop. Same camera, same drift, same "hey bro, bro, bro" underneath.

1 photo
11 seconds
9:16 vertical
30 fps
Original sound

Eleven seconds, one handheld take, the sound of the original reel underneath.

What you get back

One continuous handheld shot, no cuts. The float drifts toward the drop, a hand pushes it back, you look up from the phone, and then it glides off down the pool until it is a pink dot against the skyline. Your photo replaces the person on the float; the pool, the edge, the float and the hand stay exactly where they are.

The rider

One photo, face clear and front-facing.

Three frames from the clip

The drift, the save, the wide shot. Only the person on the float changes.

0–3 s: the float drifts toward the open edge, phone up, nobody watching

3–5.5 s: a hand comes in from behind the camera and pushes it back

7–11 s: the float glides away and shrinks into the skyline

What is the flamingo float AI trend?

Flamingo Float AI is the photo-to-video version of an 11-second reel that took over TikTok and Instagram in the first days of November 2026. A man lies back on a giant pink inflatable flamingo at the very edge of a rooftop infinity pool above Dubai, scrolling his phone, apparently unaware that the float is about to drift over the drop. A hand comes in from behind the camera, pushes the float back, and a voice says "hey bro, bro, bro" — then a quieter "thank you".

Two reels sit behind the trend. On 29 October 2026 the AI filmmaker Ilia Prudnikov (cain.aii on Instagram) posted a clip labelled AI content: a float tipping over the glass edge of a Singapore skyscraper pool and rocking back. On 2 November 2026 Younes Zarou posted the Dubai version with the hand and the thank-you, and it took off — 2.2 million likes on Instagram and 4.6 million views on TikTok within a day. Half the comments under it ask whether it is real.

That is the whole joke, and it is why the clip works as a template. The drop is the setup and the hand is the punchline. Nothing has to be acted, because the float, the edge and the skyline are already in the reference; the only variable is who is lying on the float. One photo covers it.

How to make the flamingo float video

One photo and one button. There is no prompt box, because the shot is already decided.

1
Add one photo
A clear, front-facing photo of one adult. The face, hair, skin tone and clothes come from the photo; the pool, the float, the hand and the drift all come from the clip.
2
Pick a resolution
480p, 720p or 1080p. The credit cost sits on the button before you pay; 1080p is an AI upscale of the 720p master, not a second render.
3
Download and post
11 seconds, 9:16, MP4, no watermark. The original sound is already under the picture — keep the platform's AI label on when you post.

The one seat on the float

Every version of this trend has exactly one person on the float, and that is the whole casting decision. There is no second seat: at this framing the head is roughly 25 pixels wide at the start of the clip and under 10 by the end, so splitting the frame between two faces would leave neither readable. One photo, one rider.

The rider

Lying back on the pink flamingo float, phone in hand, drifting toward the edge.

Replaces the man on the float. The hand, the pool and the skyline stay.

What photo works on the float

The face is small here, so the photo has to be sharp and close. This is the one effect on the site where a tight headshot beats a full-body shot.

Works

One person, facing the camera, head and shoulders clear
Sharp and evenly lit — daylight or a window in front of you
The face fills a good part of the frame; the render shrinks it fast
Any background: the pool, the float and the city all come from the clip

Does not work

Sunglasses or a hat that hides the eyes and the hairline
A heavy beauty filter, or a screenshot of a screenshot
Two or more people in one photo
A face turned away, half in shadow, or too small in the frame

Same portrait rules as every other effect on the site. Tiny Desk AI photo requirements

What a flamingo float video costs

Eleven seconds against an eleven-second reference is the shortest full-scene render on the site, so it is the cheapest trend clip we sell: 5 credits at 480p, 7 at 720p, 12 at 1080p. One $6.99 pack covers a 720p video.

Cast	480p	720p Best value	1080p
1 personOne photo, one rider	5credits	7credits	12credits

Priced per finished video, never per seat. 1080p is an AI upscale of the 720p master.

Credit packs: 7 credits for $6.99, 16 for $14.99, 50 for $44.99. Credits last 12 months. Pricing

Flamingo Float AI FAQ

Keep reading

The rest of the site explains the format, the photo rules and the other effects.

One photo, eleven seconds, one very bad idea

Use a photo where the face is sharp and close — it is the only thing the render has to get right.

Make my flamingo float video