Form:Privacy-FirstHealth.

The Problem

Most fitness apps share the same fundamental flaw: they require an account before you can do anything. Your data lives on their servers, behind their subscription, subject to their privacy policy. For users who are privacy-conscious, or who simply don't want to commit before trying, the friction is a dealbreaker.

The second problem: fitness apps are cloud-dependent. No internet? No access to your history.

Architecture: Local-First

The core architectural decision was to make IndexedDB the primary data layer, not a remote database. This means:

• The app loads and works instantly — no auth, no network request, no spinner
• All data is stored on the user's device by default
• Supabase sync is opt-in, triggered only when the user creates an account

I used Dexie.js as the IndexedDB wrapper. It provides a clean, Promise-based API and live query hooks that re-render React components when data changes — similar to how Supabase's real-time works, but entirely offline.

The Dual-Mode Pattern

Guest Mode  →  IndexedDB only  →  Zero account friction
Sync Mode   →  IndexedDB + Supabase  →  Multi-device access

Critically, switching from Guest to Sync Mode migrates local data to the cloud — nothing is lost. Users can try the full product before committing.

Key Features

Weight & BMI Tracking

Daily weight logs stored locally with automatic BMI calculation. The trend chart uses Recharts with a custom gradient fill and reference lines for goal weight and rolling average.

Smart Goal Forecasting

The forecast engine calculates a daily rate of change from the last 30 days of data and projects a goal completion date. Edge cases handled explicitly:

• Not enough data (< 7 days): prompts user to keep logging
• Plateau (< 0.01kg/day change): flags as stagnant
• Off track (trending away from goal): flags as off track
• Unrealistic forecast (> 730 days): shows "Long Term Goal" rather than a specific date

AI-Driven Macro Targets

Macro targets are calculated on-device using the Mifflin-St Jeor BMR formula — no third-party nutrition API required. Inputs: weight, height, age, gender, activity level, fitness goal. Output: daily calorie target, protein, carbs, and fat in grams.

Body Measurements

Waist, chest, hips, and arms tracked over time — because scale weight alone is a poor measure of body composition change.

Google Fit Integration

Optional sync with Google Fit via the Google Fit API, so Form fits into an existing health data ecosystem rather than replacing it.

CSV Import / Export

Full data portability. Users can import historical data from other apps (if exported as CSV) and export their Form data at any time. No lock-in.

What I Learned

Building local-first forced me to think about data ownership as a design constraint, not just a technical one. Every schema decision had to account for: what happens when this data needs to sync? What happens when there's a conflict between local and remote?

The Guest/Sync dual mode also taught me that reducing friction at the top of the funnel is more valuable than adding features. The single biggest driver of retention in early testing was the ability to start tracking without creating an account.

Privacy-first doesn't mean feature-poor. Guest Mode plus offline-first architecture delivers a premium UX with zero account friction.