Skip to content

Habit Tracker

Build a local-first habit tracker focused on bottle pickup tracking, habit logging, dashboard analytics, and later camera-based event detection.

  • Frontend: React app for dashboard, logging, and chat
  • Backend: FastAPI for APIs, business logic, and AI endpoints
  • Database: Postgres as source of truth
  • Vector store: pgvector inside Postgres for embeddings
  • AI orchestration: LangGraph for multi-step AI flows
  • Future vision service: separate local service for camera or YOLO-based event detection

The app is product-first, not RAG-first.

The main data is structured application data:

  • habits
  • habit logs
  • bottle events
  • notes

Later, AI and vision features will build on top of this data.

  • manual habit logging
  • bottle pickup logging
  • dashboard summaries
  • basic AI chat over structured data
  • semantic retrieval over notes
  • thread-based AI assistant
  • vision events from camera
  • posture tracking