Pawvis: touch-free hand & voice control for your Mac
Skip to content
Download
Touch-free gesture & voice control for macOS
A mouse made of thin air.
A Mac that answers to its name.
Pawvis reads your hand through the webcam. Raise it in the air and the<br>cursor follows, dip a finger to click, or say<br>“Pawvis, open Safari” out loud.<br>No mouse under your palm, no trackpad, nothing to touch. Your camera<br>never sends a frame anywhere.
Download for macOS
View on GitHub
Signed & notarized · macOS 14+ · Free & open source (MIT)
Gestures
Your webcam is the sensor.
The camera already built into your Mac watches your hand and tracks its<br>21 joints many times a second, using Apple's Vision framework. Nothing to<br>wear, nothing to hold, no depth sensor, no hardware to buy. Sit down, hold<br>up a hand, and your Mac can already see it.
Tracking watches whenever it's on, but the cursor only follows once you<br>show an open hand : all four fingers up, thumb free. Close<br>a fist for a moment, or leave the frame, and it parks. A hand that's merely<br>visible, resting or typing or gesturing, never drags the cursor around.
Move
Hold your hand open, fingers up, and move it. The cursor rides your palm.
Click
Dip your index finger, like tapping a mouse button. A quick release is always a clean click.
measured against your middle finger, so tilting your hand can't click
Right-click
Dip a second finger. Hold it down to right-drag.
pinky by default, configurable
Drag & hold
Keep the finger down and move. Deliberate movement starts a drag immediately; a short window protects quick clicks from becoming accidental drags.
the window length is a slider
Double & triple click
Tap again quickly in the same spot. Chained clicks land where the first one did.
Scroll
Fold your middle and ring fingers in, index and pinky up, then move your hand up and down. The cursor parks while the pose is held.
toggle it, or invert the direction, in Settings
Reach adapts to distance
Auto mode sizes the tracking area from how big your hand looks, so the whole screen stays reachable up close and far away, with your fingers still inside the camera frame. Manual mode gives you a fixed area and a slider.
“Show an open hand. The cursor is yours.”
On screen, a claw rides your palm: open while pointing, purple while the<br>left button is held, blue for the right, ringed in light blue while<br>scrolling, faded while control is parked. A ring tightens as your click<br>forms, a pulse confirms every click, and small dots mark each detected<br>fingertip. The menu bar icon opens a live status panel (hands seen,<br>control armed, voice state) plus a Gesture Guide that walks through every<br>move. Everything above is tunable in<br>Settings → Gestures .
In motion
Eight seconds, no hardware.
One hand, one built-in camera, and a cursor that goes where she points.
Pawvis moving the cursor by hand, live.
Voice beta
Talk to your Mac<br>by name
Switch it on in Settings → Voice (Beta) , press<br>Start in the menu bar, and say the wake word. Every command<br>begins with it, so anything you say that isn't meant for Pawvis is ignored:<br>never acted on, never typed, never put on screen.
out loud
Pawvis, go to github.com
Pawvis, type good morning
Pawvis, press command shift T
Pawvis, open Notes
Pawvis, switch to Chrome
Pawvis, scroll down a page
Pawvis, click sign in
Pawvis, stop listening
The wake word is configurable, and mishearings are tolerated. Speech<br>recognition is Apple's on-device engine: private, free, no API key, no<br>cloud. Anything that isn't a set phrase goes to on-device<br>Apple Intelligence, which maps your words to an intent and grounds<br>screen commands like “click sign in” against the area around<br>your pointer, widening to the whole screen only if the target isn't<br>nearby. That part needs macOS 26; everything else works without it.
optional · off by default
Or hand the whole command to an agent
Point Pawvis at an agent CLI you already have installed,<br>Claude Code or Codex , and it stops<br>parsing your speech and starts delegating it. Say<br>“Pawvis, clean up my downloads folder”<br>and everything after the wake word is piped to the agent to carry out<br>with computer use, running headless in the background.
A panel streams the agent's output live while it works, with a Cancel button.
Running sessions stay listed, and cancellable, under Settings → Voice.
The outcome, success or failure, always flashes in the on-screen capsule.
“Pawvis, stop listening” stays local, so you can always shut it off instantly.
Read this part twice: the agent does not ask.<br>Claude Code and Codex are launched with their own permission<br>prompts turned off, so nothing pauses to confirm anything. Whatever the<br>agent was handed, it carries out, with your account, your files and<br>your logged-in sessions: deleting things, installing things, running<br>shell commands, sending things on your behalf. Speech recognition is<br>not perfect, and a misheard command is...