About Facet

Inspiration

Facet came from a gap I kept seeing: people want coaching, but many tools feel too generic, too expensive, or too slow to use in real life. I wanted something that feels like a personal system, not another content feed, and gives guidance in the exact moment it is needed.

What I Built

Facet is a private, on-device AI life coaching app organized by life aspects like career, money, health, relationships, and more. In the app, users can:

• Browse specialized coaches by aspect
• Add personal context, goals, and values
• Start chatting immediately with minimal setup
• Unlock a cross-aspect "Sage" coach in Pro
• Upload personal sources and create custom coaches (Pro)

How I Built It

I built Facet as a React Native (Expo) app with a local-first architecture:

• On-device onboarding and personal context capture
• Session-based chat flows per coach and aspect
• Local memory and source handling with a unified knowledge hub
• On-device model pipeline for generation and transcription (including a model download step)
• RevenueCat-based entitlement flow for Pro features on supported native builds

Challenges I Faced

The hardest part was balancing privacy, speed, and product simplicity:

• On-device AI means heavier setup (large model download) and tighter performance constraints
• Keeping the UX minimal while still supporting memory, sources, and custom coaches
• Hardening monetization and entitlement behavior across platforms
• Reducing overlap in knowledge, memory, and source experiences into one clean mental model

What I Learned

I learned that trust and clarity are product features. If I promise private, personal coaching, every screen has to reinforce that with simple flows and transparent behavior. I also learned that great AI UX is less about adding more features and more about delivering fast, context-aware guidance that feels immediately useful.

Built With

• executorch
• expo.io
• llama
• react-native
• revenuecat