Red and Black Knights Sandbox
Fun chess knight visualization, but it's a screensaver without real gameplay.

Posture awareness via AirPods sensors, but solves micro-problem for niche audience.
Knowledge workers with headphones who sit for extended periods
Upright (posture phone app) · Nekoze (Japanese posture monitor)
I tried a few “reminder” style approaches and apps, but they didn’t stick. The problem wasn’t motivation. It was awareness. I didn’t want an app telling me what to do. I wanted something that simply showed me what was happening, in the moment, with zero friction.
So I built Headjust: a small native macOS app that lives in the notch area and visualizes head movement from AirPods (or compatible Beats). You can start and stop sessions while you work, then look back at simple session insights and trends with head movement visuals for better readability.
What surprised me after using it myself is how quickly it changed my behavior without any prompts. Just seeing the pattern made it easier to reset between tasks, notice when I’d gone “still” for too long, and make small adjustments before hours passed.
What it does:
- Live head-movement visualization in the notch area
- Start/stop sessions anytime during work
- Session analytics and history to spot patterns over time
- Lightweight and designed to stay out of the way
Public beta (TestFlight): https://testflight.apple.com/join/55JfhrPA
Website: https://headjust.app/
I’d love feedback on:
- whether the session flow fits naturally into real work
- whether the notch UI is unobtrusive (especially on external monitors)
- which insights feel meaningful, and what’s missing
Thanks for taking a look!
Fun chess knight visualization, but it's a screensaver without real gameplay.
Repurposes AirPods head-tracking for posture monitoring when nothing else does this.
Scroll-to-Focus uses a CGEventTap + AXUIElement to raise the exact window under the cursor, while Eye Tracking feeds webcam frames to Vision and blends head yaw + pupil offset for per-monitor focus — a tidy mix of low-level macOS APIs and ML-ish heuristics. Real-world details are handled too (calibration flow, sound/notification feedback), but expect friction from macOS permissions and the webcam requirement.
Claude can access real focus data via local MCP; weekly reviews and pattern spotting without cloud sync.
CoreAudio daemon, zero CPU idle, installs in one command—AirPods audio, Mac mic, always.
Webcam head-tracking for window focus without buying expensive eye-tracking hardware.