Most health apps assume the user is stable.

They assume the user has energy. They assume the user can log consistently. They assume the user remembers what happened three days ago. They assume privacy is something you add later.

PainTracker was built from the opposite assumption.

The user may be exhausted. The user may be in a flare. The user may have brain fog. The user may need records for a doctor, insurer, claim, or appointment. The user may not trust cloud health platforms, and honestly, they have reasons not to.

So the core tracker is local first. It runs in the browser. No required account. No default centralized health database. Static assets are handled by the PWA layer, while sensitive entries stay on device by default.

The design principle is simple:

A health app is not trustworthy because the happy path works. It is trustworthy when failure does not punish the user.

I am especially interested in feedback from people who build PWAs, IndexedDB apps, Web Crypto flows, offline first tools, health tech, or privacy respecting products.

The app is here: https://www.paintracker.ca

The bigger question I am working through is this:

What should "safe to fail" mean for software used by people who are already in pain, crisis, or medical uncertainty?