
Hi everyone — I built a macOS-only Obsidian plugin that lets you trigger Obsidian commands using simple hand gestures via your webcam.
I’ve already submitted it for the Community Plugins review. While it’s in the queue, I’d love feedback and bug reports so I can polish the UX before it lands in the marketplace.
What I personally use it for every day:
-
Open Graph View
-
Mic mute/unmute (handy during calls/recording)
-
Navigate back
-
(Your daily command here: e.g., “Open vault switcher” / “Create new note”)
Gestures supported (current v0.1.0):
-
Victory
, Thumb Up
, Thumb Down 
-
OK sign
, Open Palm
, Fist 
Each gesture can be mapped to ANY Obsidian command — fully configurable.
Quick disclosure (security / privacy):
-
No telemetry. No tracking.
-
All processing is local (MediaPipe WASM). No note content is sent anywhere.
-
Network: one-time downloads only (native helper + model). After setup, no requests.
-
Permissions: Camera is required. Accessibility is OPTIONAL (only for dictation/mic toggles).
-
Temp files: Uses /tmp for a frame + a few small status files and deletes them when the camera stops.
Why macOS-only?
On current Obsidian/Electron versions, getUserMedia can show a “live” stream but deliver zero video frames on macOS, so I use a small native helper (Swift/AVFoundation) to capture frames. The Swift source is included in the repo for auditability.
I’m looking for feedback on:
-
UX defaults (false triggers, dwell/cooldown settings)
-
Which gestures should be the “top 3” for a stable MVP
-
Reports across different Macs/webcams (lighting/webcam models also helpful)
Repo + demo GIF + install notes:
Issues / feature requests: GitHub Issues is the fastest way for me to track and respond.
If this post feels too promo-y, tell me and I’ll adjust — I’m mainly here for community feedback. ![]()