Before you start
Have a portrait of yourself, or another adult who has agreed to appear in the video. This version uses one person on the float. A group picture makes it harder to tell which person you intend to include.
You do not need a photograph taken in a pool. Your image supplies the person's appearance; the preset supplies the scene and action. You also do not need to film a stunt or find a rooftop location.
Set aside time to inspect the result. Generation is one part of making a useful clip. Choosing a suitable input, checking the action, and saving the file matter just as much. Video generation requires paid credits; review the displayed cost before you submit.
1. Make sure the scene matches the video you want
Watch the scene preview through to the end before uploading. Several videos share the Flamingo Float name, and their endings differ. A clip where the float is pushed back is a different setup from one where it tips over or falls away.
The scene used here follows the drift-and-save sequence: one rider on a pink flamingo float, an approaching pool edge, a hand entering to push the float back, a reaction, and continued movement across the water.
Pay attention to the camera distance as well. If you want a close portrait throughout, a wide pool scene may not fit that expectation. The setting and the movement are part of the result you are choosing.
For more detail about the fixed scene and the parts that can change, read the Flamingo Float AI template guide.
2. Pick a portrait you can recognize at a glance
Open your candidate photo at a comfortable viewing size. Can you clearly see the eyes, nose, mouth, hairline, and outline of the face? Look for even light and enough detail to recognize the person without zooming into a tiny head in the distance.
A straightforward portrait is a sensible starting point. Avoid making your first attempt with a face partly hidden by sunglasses, a hand, deep shadow, or a strong beauty filter. Those choices leave more of the appearance for the model to infer.
You can crop away empty background, but leave the head intact. Cutting off the forehead or chin does not make an image more useful simply because the remaining face looks larger.
Keep the unedited original. If you later compare two attempts, knowing which photo you used helps you make a deliberate change. The Flamingo Float photo guide includes a fuller selection checklist.
3. Upload the photo and check the selection
Use the upload control to choose your image. Follow the accepted file types and limits shown beside it; a filename extension alone does not tell you whether a file will be accepted.
Once the selected image appears, check it again. It is easy to choose the adjacent photo from a camera roll or a similarly named file on a computer. Confirm that the preview shows the intended person and that the crop has not removed something important.
If you want to change the image, do that before starting generation. Selecting a file and submitting a video job are separate steps.
If the upload is rejected, read the specific message first. Use the upload and generation troubleshooting guide when the image never reaches the ready-to-generate stage.
4. Review the cost, then start one generation
Before pressing the generation button, look at the selected image, the scene, and the displayed cost together. This is the last useful moment to catch an accidental photo selection or a scene that does not match your idea.
Reading the guide and watching public examples do not require buying a video. Creating your own video is a paid action. The free and cost guide explains the distinction between browsing, previews, and generation.
Start once, then watch for the page's confirmation or status change. Repeated clicks are not a reliable way to make a slow response faster. If you are unsure whether the first request went through, check the current state before submitting another.
The preset carries the scene instructions, so this workflow does not require you to write a prompt. A separate prompt guide covers building the idea in tools that accept your own directions.
5. Follow the status shown on the page
Give the current request a chance to finish. An upload, an accepted generation request, and a playable result are different stages. The page's actual status is more useful than guessing from how long it feels you have been waiting.
Keep the page available while you work through this first attempt. Do not assume that closing it, refreshing it, or changing devices will preserve every part of an unfinished task.
If an error appears, note its wording and the stage where it happened. A rejected image calls for a different response from a finished video that will not download. Check the troubleshooting guide before creating a replacement request.
6. Review your Flamingo Float AI video twice
Watch the first pass at normal speed. Follow the story as a viewer would: can you understand where the rider is, why the approaching edge matters, and what changes when the hand arrives? A technically complete file can still have an unclear sequence.
On the second pass, check three moments more closely:
- The opening: the person should be recognizable, with a consistent face and a clear position on the float.
- The push: the hand should meet the float in a way that makes the change of direction understandable.
- The ending: the rider and float should remain coherent as they move away, without an abrupt change of scene.
Listen to the actual file too. Do not assume it contains the dialogue or music you remember from a social post.
Small differences in water, clothing, and background can appear in generated video. Decide which differences affect your clip before paying for another attempt. If something needs changing, the video quality guide helps you identify one correction at a time.
7. Download, open, and prepare the clip for sharing
Use the download control when the finished result is available. After the file saves, open that copy and play it. Seeing a video on the webpage does not by itself confirm that the complete file is stored on your device.
Keep the downloaded MP4 as your original. Make a separate copy for edits so you can return to the full sequence if a crop, caption, or trim removes something you need.
When editing, leave enough of the opening to establish the pool edge. Keep the hand and float visible at the push, and give the reaction room to register. A tighter edit helps only if the cause and effect remain clear.
Choose any added sound for the platform and use you have in mind. Check the platform's current options for identifying AI-generated content. The editing and posting guide covers framing, captions, sound, and the final export in more detail.
Ready to make your first clip?
Before you begin, check four things: the scene has the ending you want, the selected portrait is clear, you understand the displayed generation cost, and you know what you will inspect in the result.
Keep that first attempt simple. One suitable photo and a careful review give you a useful basis for deciding whether to keep the clip or change the input. You can add a caption and finish the sound after the movement works.