I Built an App to Fix My Hair-Pulling Habit (It's Not Fixed Yet)

I have trichotillomania — a compulsive hair-pulling condition. There's one thing about it that's hard to understand unless you've lived it: you don't notice your hand going up. You only realize after it's done — "ah, I did it again." It's less a willpower problem than a blind spot in awareness.

If the problem is not noticing, then maybe a machine that notices for you would help. That was the start of Don't Touch.

What I built

It's an Electron desktop app. The webcam tracks your hand and face in real time, and the moment your hand enters a zone you've configured — scalp, eyebrows, cheeks — it alerts you instantly. Detection zones are finely adjustable, and daily stats and streaks let you see the trend.

The most important technical decision was that all video processing stays on the device. MediaPipe Vision runs well enough in a browser-grade runtime to make that possible. This is an app that watches your face all day. If even a single frame of that left the machine, I wouldn't use it myself. Nothing is recorded or transmitted — only the detection events remain.

An honest record of whether it worked

From here on, this is a record, not a pitch.

While I worked on a laptop, it genuinely helped. The alert would fire, my hand would come down, and at minimum I became aware of where my hands were dozens of times a day. Then I moved back to a desktop — which had no camera. Same app, but an environment where it couldn't run, and so it sat abandoned for a while. The person who built the tool proved, firsthand, that tools depend on environment more than willpower.

I posted it on Reddit once. A brief flicker of attention, then quiet. Then one day, a feature request arrived in my personal email. The request was specific: more varied alert sounds, in multiple languages, and a way to add your own mp3 or wav files. They seemed to be Russian, so when I generated the new voice alerts with AI TTS, I made sure Russian was one of the languages. I shipped the update right away, replied, and even got their feedback afterward. That one email weighed more than any download count — it meant someone with the same problem exists, and something I made is actually part of their day.

Writing down the limits too

I want to be clear about something. Trichotillomania comes from an underlying anxiety mechanism. What this app catches is the surface of the behavior — not the anxiety underneath it. I think consistent use helps, but my honest view is that an app alone won't fix it.

The evidence is simple: I built it, and mine isn't fixed yet. And if the condition is disrupting your life, professional counseling and treatment — not an app — is the right first step.

I'm keeping the app around anyway. A mirror pointed at your blind spot is clearly better than nothing — laptop-era me can vouch for that.