Your photos

How to Make AI Videos With Your Own Photos

The photo does one job. It tells the clip who is standing there. The room, the mic, and the timing stay.

Two performers face each other under a microphone against an orange wall.

A useful photo is plain in the right way. One person, face large, looking toward the camera, in ordinary light. JPG, PNG, or WEBP, up to 10MB. A group shot, a heavy filter, or a hand over the mouth gives the generation too little face to hold onto.

One performer gestures toward the microphone while the other turns away. — Two people in the frame means two photos. Each face comes from its own picture.

Match the number of photos to the template. Bulin 47 asks for one. Hotel Lobby asks for two. The trio dance asks for three. The same picture in every slot makes every person look like the same person.

01
Pick a clear face
Front view, face large, no heavy filter. The eyes should be visible even when the person in the clip wears sunglasses.
02
Use one photo per person
A different picture for each slot. The clothes in the preview stay with the clip.
03
Follow the slot labels
Person 1, Person 2, and Person 3 each replace a specific spot. The template page names the spot.
04
Compare the result to your photo
Look at the face, then the hands. If it drifts, try a sharper photo and run it again.

Both performers gesture at once against the orange background. — Fast hands are where a soft photo shows up. A sharp face holds up better.

You are swapping who is in the clip. The wall, the microphone, the crowd, and the length stay. Fifteen seconds, 10 credits to start, and a failed run returns the credits.

Templates