What it captures
Reel can take the full screen, a single window, or a region you drag out. The same panel takes still screenshots, so the tool you reach for to grab a frame is the one you already have open.
Recordings arrive back in Reel rather than in a folder you have to go looking for. That matters more than it sounds: the gap between stopping a take and seeing whether it was any good is where most re-records come from.
Why the desktop portal matters
On a modern Linux session a program cannot simply read the screen; the compositor has to hand it over. Reel asks through your desktop's screen-casting portal, which is the same mechanism your browser uses for a video call.
That is why the picker shows you exactly which screen or window you are about to share before anything is recorded, and why there is no separate permission file to edit first.
Trim the take before anyone sees it
Almost every recording needs the first few seconds and the last few removed. Press Edit and the take is on a multi-track timeline: drag a clip edge to trim, split at the playhead, move clips with snapping, and undo anything.
Save the result as a project file and it reopens with a double-click, with your cuts intact.
Captions without sending the take anywhere
Screen recordings are usually walkthroughs, and walkthroughs are mostly talking. One button transcribes the speech on your own machine and produces a caption list you can correct in place.
The speech engine and its model are fetched once, about 85 MB, and after that captioning works with no network at all. There is no account, no upload and no per-minute billing, which is the part that matters when the recording shows something internal.
Export it where it is going
Encode to H.264, H.265, AV1 or VP9 with a quality preset or your own CRF. If you only need a different container, a lossless remux rewrites the wrapper without touching the picture.
One-click destinations for YouTube, TikTok, Reels, Instagram, Facebook and X pick the frame, fit and codec for you, and deliver the mix at the loudness those platforms expect instead of letting them squash it.
Recording as part of a script
Everything the editor does, the command line does too, with no window at all. Commands speak JSON on request, so a script or an agent can cut, caption and render a take without a display attached.
Project files are plain JSON, so a pipeline can build an edit by writing a file rather than by driving a user interface.
Common questions
How do I record my screen with system audio on Linux?
Open Reel, choose the capture panel, pick the full screen, a window or a region, and start recording with system audio enabled. The take opens in Reel when you stop, ready to trim, caption or convert.
Do I need OBS as well?
No. Screen recording is built into Reel and speaks your desktop's portal directly, so there is no second tool to install, configure or keep in sync.
Does the screen recorder work on Wayland?
Yes. Reel requests the screen through the desktop's portal, which is how a Wayland session is designed to share a screen, so you get the standard picker rather than a workaround.
Can I record one window instead of the whole screen?
Yes. Full screen, a single window and a dragged region are all options, and the same panel also takes still screenshots.
Can I add captions to a screen recording?
Yes, and the audio never leaves your machine. Reel transcribes the speech locally and gives you an editable caption list that burns into the export exactly as previewed.
Free and open source. Linux-first, with a Windows build too.