Flamingo float AI is an 8-second clip in which a person lounges on a giant pink flamingo float at a rooftop infinity pool edge until an unseen hand pushes the float over the drop. On videosi you make it from one photo of yourself, with no prompt and with the "hey bro, thank you bro" sound already on it.
This guide goes deeper than the FAQ on the trend page. It covers what the clip shows beat by beat, where the sound fits, how it differs from a CapCut template, and what to do when the face drifts.
Key takeaways
- The trend is a four-beat, 8-second joke: lounging, a hand pushes the float, the slide, and a drift over the edge.
- It began spreading on TikTok in early November 2026 and is still emerging; no original creator is confirmed.
- One clear photo of one adult, showing face and upper body, gives the best likeness.
- A standard 8-second video costs 7 credits, with packs from $6.90, and failed renders return credits.
- Label the result as AI-generated when you post it.
In this guide
- What is the flamingo float AI trend, in plain terms?
- What does the clip show, beat by beat?
- What does "hey bro, thank you bro" mean?
- Why an infinity pool and a flamingo?
- How is a CapCut template different from a prompt-based video?
- Why do photos of one adult work best?
- How should I label AI content when I post it?
- Is the trend over yet, as of 7 November 2026?
- Is it safe to upload my photo?
- What if the face drifts?
- What does it cost?
What is the flamingo float AI trend, in plain terms?
It is an 8-second clip with one joke. A person lies back on a giant pink flamingo float at the very edge of a rooftop infinity pool, phone in hand, with a hazy city skyline behind. A hand from the camera's point of view pushes the float, and it slides over the edge while the camera pulls back until the float is a tiny pink dot.
The format started spreading on TikTok in early November 2026. Nobody is confirmed as the original creator, and most clips were made with a CapCut AI template. On videosi you can make your own version from one photo on the flamingo float AI page.
What does the clip show, beat by beat?
There are four beats across roughly 8 seconds.
- 0 to 1.5 seconds, lounging: one person lies back on the float at the pool edge, phone in hand, skyline behind.
- 1.5 to 3 seconds, the hand reaches in: a first-person hand and forearm enter from the corner of the frame and give the float a friendly push.
- 3 to 5 seconds, the slide: the float glides toward the edge while the person looks up, braces and calls out "hey bro" in the sound.
- 5 to 7.9 seconds, over the edge: the camera pulls back and rises as the float drifts away to a tiny pink dot over the drop.
The timing matters. The joke only lands if the push, the call-out and the pull-back arrive in that order, which is why the videosi version builds the beats in rather than leaving them to chance.
What does "hey bro, thank you bro" mean?
It is the spoken line most flamingo float clips use as their sound. The person on the float calls out to the unseen friend who pushes them, and the joke is that the push goes a little too far. That is why people also call this the hey bro thank you bro trend.
It is a spoken sound rather than a song. Posts label it as original sound, and we have not confirmed who first used it. The videosi video comes with that sound already on it, lined up with the push. The mouth is not lip-synced to it, so do not expect visible speech.
Why an infinity pool and a flamingo?
The setting is a small piece of visual comedy. An infinity pool edge looks like a drop with nothing beyond it, so a push toward that edge reads as danger at once. A giant pink flamingo float is the opposite: silly, bright and harmless. Putting the two together makes the fall feel funny instead of scary.
It also suits a short clip. The viewer understands the whole situation in the first second, and the tiny pink dot at the end is an easy, repeatable ending. Older flamingo float videos on TikTok are ordinary pool fails, which is a different thing from this trend. If you like the push-over-the-edge joke, the AI couple cliff prank and the high altitude trampoline AI trend play a similar drop for laughs.
How is a CapCut template different from a prompt-based video?
A template swaps a face or a photo into a fixed edit, so you get that edit's pacing but little control. Many of the early flamingo float clips came from CapCut AI templates, and results depend on which template version you land on.
A prompt-based video model builds the clip from instructions. That gives more room, but also more ways to go wrong. In our notes on this clip, a model without an identity lock keeps the man from the reference and only borrows your hair or shirt. The pushing hand can grow a face or a second person. The float can stay put or tip at the wrong moment.
Flamingo float AI on videosi sits in between. The scene, the push and the timing are fixed, and you supply only your photo. Under the hood it uses Seedance with a motion reference, so you never write a prompt.
Why do photos of one adult work best?
The clip is built for a single adult, and the photo is the only source for your face, hair, skin tone, build and clothes. A few habits help.
- Show the face and upper body. Face-only crops make the model invent the body.
- Look relaxed and toward the camera. In the clip you look up at the push.
- Use daylight. Dark or heavily filtered photos drift the most.
- Crop to one person. Groups, collages and pets confuse the model.
Swimwear and legs may be invented if your photo does not show them.
How should I label AI content when I post it?
Say plainly that the clip is AI-generated. The videosi steps for this trend end with exactly that advice: post it with the trend sound and label it as AI-generated. Use the platform's own AI label if it offers one, and add a short note in the caption such as "made with AI from one photo".
Only use photos of yourself, or of an adult who has said yes. Putting someone on a float without their say-so is not a harmless joke, and a label does not fix that.
Is the trend over yet, as of 7 November 2026?
No. Our stage for it is emerging. We first saw it about 1 November 2026. One how-to guide counted 87 creators with an AI version in the week before 5 November, and the biggest tutorials we found were dated 4 to 6 November.
That is a very young trend, and nobody can say how long it will run. The page was published at the start and will be updated as it moves. If you want to post one, sooner is safer than later.
Is it safe to upload my photo?
Use a photo you are comfortable sharing with a service, and do not upload other people's photos without their consent. Choose a plain portrait rather than an image that shows your address, documents or children in the background.
The finished 8-second MP4 has no watermark and is saved in My creations, so you can find it again. For specifics on how photos are handled, check the privacy policy on videosi.io rather than relying on a summary here.
What if the face drifts?
Likeness is not guaranteed. The float is tiny at the end of the clip, so the face is steadiest in the first half and blurs most near the final pull-back. That is a limit of the scene, not a sign that your photo failed.
- Retry with a sharper, front-facing, daylight photo.
- Make sure only one adult is in the frame.
- Avoid filters, heavy beauty effects and sunglasses.
- Judge the result on the first four seconds, where the face is largest.
If a render fails, times out or is blocked, the credits are returned automatically.
What does it cost?
A standard 8-second video (Standard model, 720p) is 7 credits. Credit packs start at $6.90 for 10 credits, and plans start at $29.90 per month. There is no free tier. Rendering takes several minutes. You can start from the flamingo float AI page.
The short version
If you only remember one thing, remember that the joke depends on timing and a clean face in the first half. Start with a good photo, then try the flamingo float AI video and post it with a clear AI label.
