Flamingo Float AIFlamingo Float AI: Photo and Prompt Guide
Choose a clear portrait for Flamingo Float AI, understand its fixed rooftop-pool prompt, and review format, duration and credits before creating.
A good starting photo gives the generator a clear person to work with. Before you choose a file, decide which image shows the face most naturally. You do not need a pool, a flamingo prop or a special holiday outfit in the original.
1. Compare your photos at their natural size
Open the originals rather than screenshots from a messaging app. Look for one person, even lighting and visible facial features. If the face is already soft, very small or heavily filtered, choose another image instead of relying on extra sharpening to recover details.
A practical check is whether you can see the eyes, nose and mouth clearly without zooming far into the photo. Avoid hands, hair, sunglasses or text covering important features. These are input-selection tips, not a guarantee that the resulting video will preserve every detail.
2. Keep the input simple
Choose a photo with just one person. A group image makes the intended subject ambiguous; a collage adds several competing faces or scene fragments. Crop away unnecessary borders or overlaid text while leaving the head intact.
The supported formats are JPG, PNG and WebP, with a 10 MB limit. Export a normal image file rather than renaming another format to make its extension look supported. After choosing it, check the preview and filename before you continue. Selecting a photo previews it on the page; the tool uploads it privately only after you confirm generation.
3. Understand the prompt the tool uses
The current generator uses a fixed scene prompt. There is no custom-prompt text box. Its direction asks for the person in the reference photo as the only subject, relaxing safely on a large pink flamingo float in a rooftop infinity pool, with a city skyline below.
It also asks for gentle floating movement, a slow camera pullback, natural sunlight, realistic water reflections and consistent face and clothing. It excludes text, logos and dangerous actions. These are instructions to the model, not promises that every request will satisfy them perfectly.
You do not need to paste this description anywhere. Writing words onto the input image does not create a supported prompt field and can make the portrait harder to read. Use the existing format and duration controls to shape your request.
For more about the visual composition, read the flamingo float edge scene notes. If you are planning an edit or greeting, the pool float video guide connects those settings to the final post.
4. Decide the frame and duration
Choose portrait 9:16 or landscape 16:9 before starting. Both currently use 720p output. Portrait is a natural shape for a vertical layout; landscape gives you a wider frame. Neither choice locks the output to the exact framing of a reference still.
The default is 10 seconds, with 5 and 15 seconds also available. Choose the amount of screen time your final post needs. A longer duration uses more credits at the configured rate; it does not promise a better likeness or a faster result. Read the current numeric quote in the generator and the confirmation dialog. If the cost cannot load, retry that read before submitting.
See current credit packs for the actual USD catalog and purchase terms. Pack credits and video duration are different quantities, so compare them using the cost shown for your selected request.
5. Review the result as a whole clip
Watch the full video, not only its first frame. Look at facial features, hands, clothing, the shape of the float and the water around it. Check the pool edge and background during camera movement. AI video can introduce visual changes, and a clean input does not remove that possibility.
The current output has no audio. Add any permitted music, voiceover or captions in your own editor. Save your finished clip while the download is available and keep your own copy for later edits.
A realistic rooftop scene can look like travel footage. Describe it as AI generated when sharing so viewers understand what they are seeing. The existing site images are credited style references, not verified examples generated from a supplied input photo on this site.