Vertical Video Maker

The four layouts

Facecam on top, gameplay below is the classic streamer clip. Drag a box around your webcam in the source frame, or press Find my facecam and the tool spots your face and draws the box for you; it's scaled to fill the top band (35% of the height by default), and a vertical slice of the gameplay fills the rest. Slide the gameplay position to keep the action in frame, and zoom in for small HUD games. Follow my face is auto-reframe: a 9:16 window that pans to keep the speaker in frame, which suits just-chatting, IRL, podcasts and talking-head videos. Crop to vertical takes a fixed 9:16 slice. Blurred background keeps the whole 16:9 picture in the middle with a blurred, darkened copy filling the space, which is safest when everything on screen matters.

Example: a 1920 × 1080 clip with a 35% facecam band leaves a 1080 × 1248 gameplay band. At zoom 1, the gameplay slice is 1080 px tall and 935 px wide, a little under half of the original width, positioned wherever you slide it.

How auto-reframe works

The tool looks at about four frames a second (fewer on long clips) with YuNet, a small face detector from OpenCV that runs on your device. It keeps following the same face from frame to frame instead of jumping to whoever is biggest, fills short gaps when the face turns away, and smooths the path over about a second so the crop glides rather than shakes. The face sits slightly above the middle of the frame, where viewers expect eyes to be.

Example: on a 1920 × 1080 clip at zoom 1, the window is 1080 px tall and 608 px wide. If the face is 70% of the way across, the window is centered at 1,344 px and covers 1,040 to 1,648 px.

Tips for vertical clips

Keep the clip under 60 seconds and start on the action. Your facecam box is remembered for next time, so it only needs setting once per overlay layout. Add captions afterwards with the auto caption generator, and check nothing important sits under the app buttons with the safe zone checker.