How to Make a Flamingo Float AI Video
Turn one photo into a pool-float video using a pink flamingo reference. The workflow below explains what to prepare, what the template controls and what to review before sharing.
Last updated: 2026-11-07
1. Watch the source reference
Start with the preview on the homepage. It shows a single person riding a pink flamingo float in a rooftop pool, with a city skyline behind them. The reference is 11.4 seconds long and vertical, at 720 × 1280 pixels. It is the movement and scene reference for this tool, not a generated customer example.
Watch the whole clip to understand the person’s gestures, the float’s movement and the camera framing. Your photo is intended to change the person’s identity while keeping that sequence. The template is silent; the planned output uses generated pool ambience instead of copying a soundtrack.
2. Choose one clear photo
Upload a JPG, PNG or WebP of one adult. The minimum size is 300 × 300 pixels and the file limit is 10 MB. A face with even lighting, visible hair and natural proportions gives the model more useful information than a tiny face or a heavily filtered selfie.
A front view or a slight three-quarter view usually supplies more facial detail than a full side profile. Avoid sunglasses, hands over the face and group photos. You do not need to stand beside a pool or hold a float in the source photo; those elements come from the video reference. Only upload photos you own or have permission to use.
3. Pick a video quality
480p: 50 credits, 720p: 100 credits, 1080p: 200 credits per video. The 480p and 720p options use Seedance 2.0 Fast; 1080p uses Seedance 2.5. The target format is a vertical 9:16 MP4 with generated audio. Higher resolution increases pixel dimensions but does not guarantee better likeness or smoother motion.
All qualities aim to follow the full 11.4-second reference at its original pace. Fast clips run about 12 seconds, while 1080p edits target the source length. Final timing and motion can vary. Preview the reference and choose the quality that suits your needs.
4. Submit once and follow the task
When the service is available, sign in, confirm photo permission and submit the request once. Credits are reserved for that job. The result panel shows its progress, and My videos keeps the task linked to your account.
If your connection drops, reconnect to the saved request instead of creating a new one. A submission with an uncertain provider response remains under review. A confirmed failed or rejected job returns its reserved credits once; a temporary download problem retries saving the existing result.
5. Review before downloading or sharing
Watch the completed clip from beginning to end. Check the face, hands, contact with the float, scene continuity, motion and sound. AI edits may alter details, and the exact appearance of the source photo or reference is not guaranteed.
Download the MP4 from your private history. Clearly describe it as AI generated when sharing if viewers could mistake it for real footage. This site does not provide music rights or guarantee that a social platform will promote a clip. Contact support with the task ID if something fails.