1. Watch the scene before choosing a portrait
The lead begins in a black tank top and light trousers, holding a silver cup in a busy room. Watch the close-up gestures, side turns and moment when hands partly cover the face. Those movements explain why a clear identity reference matters.
The player shows a supplied demonstration, not an EasyVideoAI result made with your photo. Background guests remain part of the scene. You are casting one foreground person, so there is no second upload for the crowd.
2. Upload one well-lit person photo
Use a JPEG, PNG or WebP image with one obvious subject. Include the face, hair and upper-body clothing. The template uses the photo for appearance and outfit, while the performance posture comes from the video.
Avoid a distant face, strong blur, sunglasses hiding most of the face or a group picture with several equally prominent people. If your chosen photo contains others, crop it to the intended person before uploading. Confirm that you have permission to use the image.
3. Select the short or full performance
The 15-second choice starts at the beginning of the supplied scene. The full option also starts at 0:00 and keeps about 29.08 seconds of picture, including the later crowd movement. Choosing the full option does not invent a longer ending.
Select 480p or 720p, then review the live quote. The full clip is charged using 30 started seconds; the short clip uses 15. Keep the original near-4:3 composition to retain the cup and gestures at the edges.
4. Check the settings and confirm once
Review the selected portrait, duration, resolution and required credits before confirming. If you need to buy credits, return to the studio and check your settings again; buying credits does not itself create the video.
After submission, follow the saved request on its result page. If a connection drops, check My videos or the last saved request before submitting again. Reopening an existing result is different from starting another generation.
5. Review the face, cup and sound before downloading
Watch the full MP4 with audio. Check that the same identity remains on the lead when they turn sideways, cover their face and reappear. Look at the fingers and cup contact, and make sure the replacement has not spread to a guest in the background.
The final workflow restores audio from the selected source segment. It does not use the portrait to clone a voice or record new lyrics. If you want a vertical post, prepare that framing after downloading and inspect the entire edit again.
Reviewed against the EasyVideoAI studio on November 5, 2026. Check the live studio for current availability and the final credit quote.