LYLU AI Diary

The Problem

Journaling is powerful for mental health, but typing on a phone creates friction. Voice removes that barrier — but raw audio isn't searchable or analyzable.

The Solution

Built a native mobile app where users speak their thoughts. Audio is transcribed via Whisper, sentiment is analyzed via Hugging Face, and entries appear in a calendar view with mood tracking. Monetized with subscriptions and rewarded ads.

Preview

Technical Highlights

• Shipped to both Apple App Store and Google Play — a real product, not just a repo
• 3-service architecture: React Native app → Express 5 backend → FastAPI Whisper service
• RevenueCat subscription management + AdMob with Pro/Free tier quota system
• Google Play Integrity + iOS App Attest for app authenticity verification
• Hugging Face sentiment analysis with daily mood aggregation
• Full production deployment: AWS EC2, PM2, Nginx, Let's Encrypt SSL, MongoDB Atlas

Tech Stack

Key Numbers