A good first take.

1. Record or import

Choose Record screen and set up your microphone and camera. The browser’s picker decides which screen, window, or tab is shared. Source audio depends on that choice and your system. A normal webpage cannot track clicks in other tabs or apps; use manual focus for web recordings.

2. Edit without changing the original

Choose a clip in the timeline. Drag its edges or use numeric start/end controls in Edit. Press S to split at the playhead. Shift-click selects multiple segments. Delete removes selected segments; “Keep gap when deleting” controls whether time is preserved.

Use Focus to add a zoom, then click the preview to set its target. Masks use source coordinates and are rendered into the exported picture. Inspect every mask before sharing sensitive content. Caption text and timing are editable in the inspector.

3. Save a portable backup

Autosave keeps projects in this browser. Browser cleanup, storage eviction, private browsing, or another device can make them unavailable. The project menu’s Download backup creates a .glidelo file containing your original media and edits. Open project imports it again.

4. Export and inspect

Choose a resolution, frame rate, format, and quality. Available encoders vary by browser and device. Export happens locally in a worker and can be cancelled. Download the finished video and play it back before sharing. Speed changes currently change audio pitch.

Keyboard shortcuts

SpacePlay / pause
Left / RightStep one frame
SSplit at playhead
Ctrl / Cmd + ZUndo
Ctrl / Cmd + Shift + ZRedo
Delete / BackspaceDelete selection
Alt + Shift + SStop recording while studio is focused; extension command when installed

When something goes wrong

Permission denied: reopen setup and allow only the devices you need. Missing media: use Projects → Relink current media and select the original file. Interrupted recording: look for Recover in Projects. Export unavailable: try another offered format on a desktop Chromium browser. Do not close the tab while exporting.

Speech transcription is experimental and downloads model files on explicit request. English and Malayalam are model-supported choices; output accuracy and offline operation are not yet device-verified. You can import SRT or VTT without downloading a model.