A split-screen podcast short stacks the people in the conversation inside one 9:16 frame, so the guest does not disappear every time the host talks. The layout that works is simple: two speakers stacked top and bottom, three as a pair on top with one full width below, four as a 2x2 grid, and a single full-height shot of whoever is talking whenever the edit cuts to a close-up. This guide covers each layout, what your recording needs, how to build it by hand, and how Skapo does it automatically in about 10 minutes for a whole episode.
Why split-screen beats a single crop for interviews
Cropping a 16:9 conversation to vertical keeps about a third of the picture. With one person in that third, every question, laugh and reaction from the other person happens off screen. For a solo monologue that is fine. For an interview, the reaction is often the reason the clip works, and a viewer who sees one face talking to nobody scrolls on. Split-screen keeps both.
Podcast split-screen format: top and bottom layouts for 2, 3 and 4 people
| Speakers | Layout | When it works best |
|---|---|---|
| 1 | Full height, following the speaker | Solo episodes, and every close-up in a multi-camera edit |
| 2 | Stacked top and bottom, one per half | The standard interview: host and guest across a table or on a call |
| 3 | A pair on top, one full width below | Panels and co-hosted shows with a guest |
| 4 | 2x2 grid | Roundtables; beyond four, faces get too small to read on a phone |
What your recording needs
- At least one shot where everyone is visible together. A wide studio shot, or a gallery view with everyone in one row, as Riverside, StreamYard and Zoom show two or three people. That shot is what tells you where each person sits. A 2x2 gallery grid is not split yet, so for remote recordings pick the side-by-side layout.
- Close-ups are fine. Most studio edits cut between a wide shot and solo close-ups. Use the split while the wide shot is on, and a single speaker while the close-up is on.
- Only close-ups, never together? Then there is nothing to stack, and the honest layout is a single full-height shot that follows whoever is talking.
- Already vertical? If you recorded in 9:16, there is no width to split; keep the shot as it is.
The three mistakes that make split-screen look broken
- The empty chair. Keeping the split on while the edit shows only one person, so one slot shows a background or a stranger’s shoulder. Switch to a single speaker for those stretches.
- The jumping frame. Re-centring each slot every time someone leans in or turns. Fix each person’s framing for the length of the shot and let small movements happen inside it.
- Flickering between layouts. Switching to a single speaker for a one-second reaction cutaway and straight back. A short cutaway should not break the split; only a sustained single-speaker stretch should.
Doing it by hand
In Premiere Pro, DaVinci Resolve or CapCut: put the clip on two tracks, crop each to one speaker, scale each to 1080x960 and position one above the other on a 1080x1920 sequence. Then go through the clip and replace the split with a single full-height speaker wherever the edit cuts to a close-up, add word-level captions above the bottom fifth of the frame, and export. Expect 15 to 25 minutes per clip, so two to three hours for a week of clips from one episode.
How Skapo does it
Paste a YouTube link or upload the episode. Skapo finds the moments worth posting, and for every clip it works out who is on screen together: 2 to 4 speakers get the layouts above while they share the frame, and when the edit cuts to a close-up the short switches to whoever is talking. A brief cutaway does not break the split. In a two-person split the person talking takes the top half, and inside each slot the framing is fixed for the shot, so nothing jumps when someone leans in. Captions are timed per word, kept clear of the platform buttons, in 22 tuned languages. The whole set is usually ready in about 10 minutes, about 20 for a two-hour episode. The details are on the split-screen page.
What it does not do: put a second, unrelated video under your speaker (the “gameplay below the podcast” format). Split-screen in Skapo is always the people in your own recording.
Try it on your own episode
The free plan includes 120 minutes of video with watermarked clips, which is enough to see how your show looks stacked. Start on the podcast page, or read the full podcast-to-shorts workflow for the steps before and after the framing.