KlipKap Pro documentation

System requirements

There is no Linux build. A build is in development, but nothing on this page describes software you can run on Linux today.

Recording, editing, transcription and export all run on your machine. The only network calls KlipKap makes are downloading the app and activating a licence, so an editing session works with the connection off.

Signing in

The sign-in screen appears on first launch and after signing out. Sign-in happens in your browser, not in the app.

Sign In:

Opens the authentication page in your default browser.

Finish the login flow there. KlipKap detects it automatically and returns you to the editor. Nothing to copy back.

First launch

Once signed in you land on the start screen, which does two things.

Start a new recording:

The record button, or File → New.

Open an existing project:

File → Open Project…, or pick one from File → Open Recent.

Choosing what to capture

Click Record, or File → New. The source picker opens with three tabs.

Audio

Both audio sources are set in the source picker, before the recording starts. They arrive in the editor as separate tracks with independent volume, so a loud notification does not force you to duck your own voice.

Microphone:

Pick your input device from the dropdown. An 18-bar meter shows the live level beside it.

System audio:

Captures what your computer is playing (app sounds, music, a video you are reacting to) alongside your voice.

Webcam

If a webcam is detected it is captured automatically alongside the screen, as its own overlay track. Position, size, shape and zoom are all decided later in the editor, so nothing about the framing is locked in at record time.

Running the recording

The KlipKap window hides itself when recording starts, so it never appears in your own capture. A tray icon takes over: the menu bar on macOS, the system tray on Windows. It blinks while recording is live.

Pause:

Suspends the recording. Resume picks up where you stopped.

Stop:

Ends the session and opens the editor with the footage ready.

Layout

The preview is not an approximation. Changes made in the properties panel appear in it immediately, and the export renders the same frame, which is why almost no step in this manual asks you to export to check something.

The properties sidebar

Nine icon tabs: Props, Style, Tune, Cursor, Webcam, Subs, Mask, AI and Export. Props changes with whatever is selected on the timeline; the rest are global. Webcam appears only if a webcam was captured with the recording.

Playback

Play / Pause:

The play button, or the spacebar.

Step / Skip:

Jump one item forward or back, or straight to the start or end of the timeline.

Stop:

Returns the playhead to the beginning, or to the in-point if one is set.

Loop:

Repeats playback between the in and out points instead of stopping at the out-point.

Playhead:

The vertical red line. Drag it to scrub anywhere in the recording.

Zoom and navigation

Zoom:

The zoom slider and buttons, or hold Shift and scroll to zoom around the pointer.

Fit to window:

Scales the whole timeline to the current viewport.

Scroll:

A plain mouse wheel pans horizontally.

The in / out range

Two handles sit on the timeline ruler. Drag them to mark the portion that will be exported. Everything outside is ignored, and playback loops inside it.

Track channels

Editing items

Add:

Click empty space in a track at the position you want, or use the + at the start of the track.

Move:

Drag an item horizontally to a new time.

Resize:

Drag either edge to shorten or extend it.

Select:

Click it. Its controls load in the Props tab, and inline action buttons appear on the item.

Cut:

The scissors splits the selected item at the playhead into two segments.

Merge:

Joins a segment back onto the next one.

Delete:

The ✕ on a selected item removes it.

Cuts

Items on the Cut track mark stretches that disappear at export. Video, audio, subtitles and effects all skip over them together, so a cut never leaves the captions a second out of step with the voice.

Add them by hand, or generate them from the AI tab (see AI auto-cut below).

Props

Contextual. Select something on the timeline and its own controls appear here.

Audio item:

Mute, volume from −60 dB to +60 dB, auto-trim silence, AI Denoise (Pro), and the voice-processing sliders described under Pro features.

System Audio item:

The same controls, with mute and volume independent of the microphone track.

Zoom item:

Zoom factor from 1.0× to 3.0×. Turn on Manual zoom position to lock the camera on a fixed area with the mini-canvas crosshair instead of following the cursor.

Mask item:

The position and size of each rectangle in a censor mask, if you would rather type them than drag them.

Subtitle item:

Edit the phrase text directly. Word-level timing re-syncs on its own, so the highlight stays in step.

Style

What sits behind the recorded window. The setting is global and updates live in the preview.

KlipKap backgrounds:

Eight supplied backgrounds: Ethereal, Obsidian, Axiom, Vanta, Lumen, Novae, Flux and Zenith.

macOS wallpapers:

The stock macOS wallpapers from Big Sur through Tahoe, in light and dark variants.

Upload image…:

Any picture from your own machine.

Colour presets:

Dark and White, plus 16 gradients: Blue, Purple, Green, Rose, Orange, Teal, Amber, Indigo, Crimson, Slate, Sunset, Ocean, Aurora, Fire, Candy and Forest.

Custom colour:

Any solid colour, from the picker.

Reset background:

Back to the default.

Tune

Camera behaviour, window appearance and colour, in three groups.

Playback: how the camera moves.

Mouse smoothness (0–1):

How fluidly the camera follows the cursor path. Higher is smoother and lags further behind.

Zoom speed:

Slow, Normal or Fast. How quickly the camera pushes in and pulls back.

Follow speed:

How fast the camera pans to keep up with the pointer.

Auto-zoom:

Generate builds zoom segments from the clicks recorded in the take; Clear All removes them. The segment count is shown underneath.

Window appearance:

Colour grading and effects:

Reset appearance returns every control on this tab to its default.

Cursor

The cursor is redrawn rather than captured, which is why all of this is still adjustable after the recording exists.

Style:

Twelve pointers: Arrow, White, Peak, Doritos, Cloud, Diams, Bubble, Map, Forest, Tree, Melon and Stone.

Cursor size (16–128 px):

How large the pointer is drawn, independent of your system setting.

Cursor shadow (0–1):

Drop shadow under the pointer, which is what keeps it legible over a light interface.

Cursor tilt:

Leans the pointer into its direction of travel and eases it upright when it stops.

Hide cursor if not moving:

Fades the pointer out once it has been still, and back in when it moves.

Hide cursor:

Removes the pointer from the preview and the export entirely.

Webcam

Present only when a webcam was captured with the recording.

Enable / Disable:

Turns the overlay on or off.

Position:

A 3×3 grid with eight positions: the corners and the edge midpoints.

Size (5–40%):

Overlay diameter as a percentage of the video height.

Zoom (1.0–3.0×):

Digital zoom into the camera feed itself, to crop a cluttered room out of frame.

Mirror horizontal:

Flips the feed left to right, in the preview and the export.

Auto-zoom scale:

How far the overlay shrinks while the camera is zoomed in, so it stays out of the way of the thing you are pointing at.

Shape:

Circle and square shortcuts, plus a border-radius slider from 0 (square) to 100 (full circle).

Shadow (0–1):

Drop shadow behind the overlay.

Border width and colour:

The ring around the feed.

SubsPro

Subtitles are generated from your audio, on your own machine. No audio is uploaded, there is no API key and there is nothing metered per minute. Transcription is a Pro feature; the styling controls below apply once phrases exist.

Transcribe (Pro):

Runs locally with a progress bar. Phrases land on the Transcription track when it finishes.

Show subtitles:

Renders them, or does not.

Progressive word highlight:

Karaoke style. Each word stays lit once spoken instead of dimming back down.

Inactive opacity (0–1):

How dim the words you have not reached yet appear.

Keep short phrases visible:

Holds the previous phrase on screen at low opacity until the next one starts.

Font size (12–64 px):

Text size at 1080p.

Background opacity (0–1):

Opacity of the pill behind the text.

Border radius (0–32 px):

Corner rounding on that pill.

Bottom offset (1–40%):

Distance from the bottom edge of the frame.

Mask

Rectangles that heavily blur whatever is behind them: passwords, API keys, customer names, faces, anything that should not survive the export.

Enable:

Turns masks on and adds a Mask track to the timeline.

Add rectangle:

Drops a new blur rectangle. Drag and resize it in the preview, or type its position and size in the Props tab.

Masks stay locked to the content underneath them, so a mask holds its target through a zoom and a pan.

Use the + on the Mask track to add a mask that covers a different stretch of time.

AI

Two tools, both optional.

AI auto-cut finds dead air, stutters and filler words and adds them to the Cut track. It needs word-level timing, so transcribe the audio first.

Minimum silence length:

Only silences at least this long are cut.

Silence sensitivity:

How far below the recording's usual loudness counts as silence.

Detect & add cuts:

Runs the pass. Re-running never stacks duplicates; overlapping cuts are merged. Clear All removes every cut.

MCP Server (Pro) lets an external AI client connect to KlipKap while it is open: Claude Code, Gemini CLI, Codex, Cursor, Windsurf, LM Studio and anything else that speaks the Model Context Protocol. The client can read the timeline, the transcript and the cuts, and edit them: adding or removing cuts, rewriting subtitle text for a translation, and so on. Changes appear live in the timeline. Toggling it on gives you a ready-to-paste setup command for your client.

Export

An Output box states the final pixel dimensions and the duration before you commit to the render.

Step by step

Set the in / out range on the timeline so the export covers the take and nothing either side of it.

Open the Export tab.

Choose format, renderer, resolution, frame rate, aspect ratio and quality.

Browse to an export directory.

Optionally type a filename.

Click Export Video. A progress dialog reports the render.

Click Reveal in Explorer when it finishes to open the containing folder.

Vertical and social formats

Switch the aspect ratio to 9:16 for TikTok or Reels and the auto-zoom system reframes the content to fit the vertical canvas. The camera is already tracking what matters, so it has something to reframe around rather than centre-cropping blindly.

Keep GIF exports under about ten seconds. The format has no interframe compression worth the name, so file size climbs fast.

The watermark

Free exports carry a KlipKap watermark. Pro exports do not, and there is no step to perform.

Auto-save

KlipKap saves the project two seconds after any change. Saving by hand is only necessary when you want the file in a particular place.

File operations

Side by side

The free tier is permanent, not a trial. Recording, the timeline, auto-zoom, the cursor styling, masks, backgrounds and the appearance controls are all in it. The licence gates the export ceiling and the AI features, not the editor.

Audio cleanup

Found in the Props tab with an audio item selected. Everything here is applied at export rather than written into the recording, so the original audio is always recoverable.

AI Denoise (Pro):

Removes background noise with a neural model, in Standard or High Quality mode. Revertible at any point.

Noise suppression:

An adaptive FFT denoiser. Drag for more of it.

Bass boost:

Low-shelf boost at 150 Hz, for warmth and body.

Voice clarity:

Lifts presence and air, which is what makes speech easier to follow on laptop speakers.

Voice boost:

Compression, a presence lift and gain together, for a louder and punchier voice.

Auto-trim silence:

Detects silent stretches on the track and removes them.

All shortcuts

While recording

Pause instead of Stop. Segments are stitched automatically and the viewer never sees the seam.

Test the microphone meter before every session, not just the first one. Input devices change without telling you.

Record at the resolution you intend to export. A 1080p export from a 1920×1080 window is sharp; the same export upscaled from a small window is not.

While editing

Set the in/out range first so every preview loops over the part you are shipping.

Generate auto-zoom from your clicks, then delete the two or three segments that landed somewhere uninteresting. Fixing a generated set is faster than building one.

Transcribe early, then run AI auto-cut. Stripping dead air and filler in one pass and tidying the Cut track by hand beats hunting for pauses.

Mask anything sensitive the moment you notice it. The blur tracks the content through zooms, so it does not need re-checking later.

A macOS wallpaper or a gradient with a little corner radius and shadow gets you most of the way to a polished look before you touch anything else.

While exporting

1080p / 30 fps / Max is the right default for YouTube and for embeds.

Use the GPU renderer if you have a Pro licence. The output is identical and the render is faster.

For Reels or TikTok, switch to 9:16 and let auto-zoom reframe rather than cropping the export afterwards.