How to Create an AI Video from a Photo
A practical guide to choosing a video template, preparing a portrait, selecting output settings and downloading your finished AI video.
Start with the scene you want
Open NEON Duet to review its scene, photo requirements and generation settings. The artwork describes the visual concept; published examples, when available, are labeled separately.
Prepare a clear portrait
Use JPG, PNG or WebP, at least 340 × 340 pixels and no larger than 8 MB. A sharp face, even lighting and an unobstructed expression give the model a clearer starting point. Upload only photos you have permission to use. For a duet, use two distinct portraits, one for each person.
Choose settings before you generate
NEON Duet uses EvoLink Seedance 2.5, 480p landscape video and a 15-second reference performance. Generation availability depends on the EvoLink connection and template assets.
Continue creating while the job runs
Sign in, add your photos, confirm permission and submit. The saved task runs in the background, so you can leave the page. Personal center → My videos shows its current status, including queued, preparing, rendering and completed. A completed video can be previewed and downloaded.
Download within the storage window
Completed video output is temporarily stored for 24 hours from completion. Download a copy during that window. After expiry the task remains in your history, but its playback and download are unavailable. This video retention window does not describe the retention of source photos.