Life Journal — AI Journal (Mobile, MVP)
Startup MVP built with a business partner. Mobile app in Expo for iOS/Android, currently testing on TestFlight and Android. Backend in FastAPI deployed on Railway. Includes authentication, journaling features powered by RAG, image uploads, and voice recording transcription with OpenAI models. Uses Supabase for storage and vector database functionality.
Problem
Users want a journaling experience that can recall context (RAG), understand voice notes, and handle rich media (images) while staying fast and usable on mobile.
Constraints
- • Mobile-first UX with offline/low-latency expectations
- • Reliable media uploads (images/voice) and processing
- • RAG pipeline needs consistent embeddings + storage
- • Secure auth + user data boundaries
Solution
Built a cross-platform mobile app with Expo and a FastAPI backend deployed on Railway. Implemented auth, media uploads to Supabase, voice-to-text transcription, and a RAG pipeline for context-aware chat/journal insights using OpenAI models and vector storage.
Key Features
RAG Journal Assistant
Context-aware AI that references your past entries.
Voice Transcription
Record voice notes and transcribe them via AI models.
Image + Media Support
Attach images to entries and store them securely.
Cross-Platform MVP
Single codebase for iOS + Android with fast iteration.
Tech Decisions
Expo for Speed + Cross-Platform
Expo enabled fast iteration and a single codebase for iOS/Android during MVP stage.
FastAPI Backend on Railway
FastAPI provided a clean API surface and fast development; Railway simplified deployment and environment handling.
Supabase for Storage + Vector
Used Supabase for media storage and vector database capabilities to support RAG features.
Performance & Metrics
Lessons Learned
- • Media pipelines need strong retry/error handling from day one
- • RAG quality depends heavily on consistent chunking/embeddings strategy
- • Mobile UX must stay simple even when features are complex