Device names appear once you have allowed access at least once. Until then a browser only says how many there are.
Recording this tab is the only case where a browser knows where your pointer is, so it is the only case where the zoom can follow your cursor exactly. Recording anything else still works: the zoom follows where the picture changes instead, which needs no cursor at all.
Recordings kept here
A recording is kept in this browser as soon as it stops, so closing the tab or reloading will not lose it.
Projects and checkpoints
Copies include the media and use extra browser storage. Restoring a checkpoint opens a new copy and keeps the saved checkpoint.
Transcript tools and caption style
Search and timing use the source recording. Downloaded transcripts follow the finished edit. Select adjacent lines in the transcript to merge them.
Review silence cuts
The threshold is relative to the loudest moment. Padding is kept at each end; the minimum gap is measured after padding. Review each proposed cut before applying it.
Review and queue exports
Each queue entry captures the current edit and framing. Set the output shape, crop and caption placement, then add that version. Keep this page open while exports run. Download each completed file from the queue.
Nothing recorded yet.
Drag on the picture to say what this zoom should look at. The bright rectangle is what it will show.
Drag on the picture to place what is hidden. It is burnt into the exported video, not laid over it.
Drag on the picture to place it. An arrow runs from where you press to where you let go.
Lookshape, background, angle, saved looks
Where it is going
Sets the shape and the output size together, and shows what will be cut off.
Background
Angle and arrival
The arrival is measured from where the export begins, not from the start of the recording.
Looks
The background, padding, shadow, tilt, entrance and zoom defaults, saved under a name so the next recording can match this one.
Wordstranscript and chapters
Transcript
Bring in an SRT or VTT file and the words become subtitles you can burn in. Selecting lines and pressing Cut removes those seconds from the video, which is the quickest way to take out a sentence you did not mean to say.
The words are worked out on this machine. Doing that needs a model, which is fetched the first time you ask and then kept, and is the only thing on this site that touches the network. Your recording stays here.
Chapters
Press M while recording to mark a moment. These are what you paste under a video.
Motion, sound and outputautomatic zoom, camera, audio, format
Automatic zoom
These are the defaults the automatic zooms are built from. To change one zoom, click it on the timeline instead.
The cursor and the ripples need a pointer track, so they only appear when you recorded this tab.
Camera and output
Roughly . Every frame is drawn and encoded one at a time, so a long recording takes a while and comes out exactly as the preview shows it.