One take. Three layers.
Pick your screen, camera and mic from the device pickers — mic level meter included. Compose the camera as a bubble, side-by-side, screen-only, or camera-only. Pause and resume mid-take; chunks upload while you record, so a dropped connection never costs a take. Fifteen-minute auto-stop keeps takes sane. In a hurry? Quick mode burns the bubble into one file.
Layers, not a flat file.
Layers mode saves the camera as a separate track. Your take opens in the editor as a real composition — re-layout the camera bubble, crop it, push it full-frame, reframe the whole thing to 9:16. Change your mind after recording, because the layers never fused. The editor deep-links straight to the take (?open=); quick mode is there when one file is genuinely all you need.
Every word, searchable.
Finalize the take and faster-whisper transcribes it word by word — proxies (720p/540p), thumbnail and scrub-rail filmstrip follow automatically. Search "the pricing demo" and click the moment — the editor opens at that second. Recordings are first-class library citizens.
Ship it the same tab.
Karaoke captions, brand accent, 9:16 reframe with safe zones, −14 LUFS loudness, guest review link with frame sketches. The whole pipeline — one product, no round-trips.
Open a tab.
Start talking.
Your recording becomes a searchable, editable project before you finish your coffee.