Make an AI kiss video from one photo of you both

Two separate photos are the default because most people do not have a good picture of the two of them together, or want to put together two people who were never photographed side by side. If you do have that picture, it is usually the better input. The light, the scale of the two faces and the background already agree, so the model has nothing to reconcile and the clip looks like the moment it came from.

It is also the right mode for re-creating a real memory: a wedding photo that becomes a kiss, a holiday selfie that becomes a hug, the first photo you ever took together turned into an anniversary clip.

In one-photo mode the app scales your photo to fill the whole frame, keeping its proportions, and trims whatever sticks out. That happens in your browser before anything is uploaded. It means the frame shape decides the crop: a wide landscape photo dropped into the tall 9:16 frame loses its left and right edges, and a tall portrait photo in the 4:3 frame loses its top and bottom.

So pick the shape that matches the photo, not the one that matches where you will post. If both of you sit near the edge of a landscape picture, choose 4:3, or crop the photo yourself first so the two faces are near the middle.

The model starts from the pose in your photo, so the closest styles to that pose look the most natural. Two people already cheek to cheek move into a Kiss, a Cheek kiss, a Selfie or Laugh together with very little invented movement. Two people standing a step apart suit a Hug, Hold hands or Slow dance, where stepping in is part of the action.

Every style is listed, with what happens in each one, on the video styles page. The price depends only on the length, so trying a second style on the same photo costs one more generation and nothing else.

Older couple photos work well in this mode, because what matters is focus and face size, not age. Scan or photograph the print flat, in daylight, without glare, and crop away the white border.

What a single old photo cannot do is fix a face that is tiny or blurred in the original. If the two of you are small figures in a landscape, the result will be soft. In that case two separate close-up photos, even from different years, usually give a better clip than one distant shot.

The photo is read, resized and framed on a canvas inside your browser, and that single composed frame is what is uploaded for generation, not your original file. There is no account to create. The privacy and safety guide and the privacy policy cover what is kept and how to ask for deletion, and is AI Kiss safe? answers the payment questions.