Zainab4626/habit-bloom-464

🌱 HabitBloom — AI-Powered Habit Tracker

A full-stack habit tracking application with AI coaching, voice summaries, and data-driven insights — built with React, TypeScript, and Supabase.

Live Demo →

---

✨ Features

Core

• User Authentication — Email/password signup & login with session management
• Habit Management — Create custom habits with color coding, icons, and weekly targets
• Dashboard — Real-time progress rings, streak tracking, and weekly completion metrics
• Habit Logs — Full history view with filtering and status management
• Mobile Responsive — Bottom navigation on mobile, sidebar on desktop

AI-Powered

• Habit Coach Chat — Ask questions about your patterns and get personalized, data-driven advice powered by an LLM (Gemini)
• Voice Summary — One-tap weekly summary generated by AI and read aloud via browser Text-to-Speech
• AI Insights Lab — Structured report with habit health scoring, risk detection, and experiment-driven recommendations generated from your last 6 weeks of logs

---

🔬 Insights Lab — AI Engineering Deep Dive

The Insights Lab is a production LLM analytics pipeline that turns raw behavioral logs into a structured coaching report.

Pipeline:

1. Data aggregation — Supabase Edge Function pulls 6 weeks of habit logs, computes a deterministic baseline consistency score, and builds per-habit completion summaries.
2. Structured prompting — Google Gemini (via Lovable AI Gateway) is prompted with a strict JSON schema covering health score, trend classification, at-risk habits, key patterns, and hypothesis-driven experiments.
3. Output validation — Server-side JSON extraction with fallback parsing, type coercion, and field-level guards. Client validates with Zod before rendering.
4. Caching — Reports cached in localStorage to avoid redundant LLM calls and reduce latency on repeat views.

AI engineering concepts demonstrated:

• Schema-constrained generation (JSON-only output, no markdown leakage)
• Grounded prompting (LLM only sees user data + computed baselines, no hallucinated context)
• Defensive parsing for non-deterministic model outputs
• Hybrid deterministic + LLM scoring (baseline math anchors the model's health score)
• Hypothesis → action → success-metric framing for actionable AI recommendations

---

🛠 Tech Stack

Layer	Technology
Frontend	React 18, TypeScript 5, Vite 5
Styling	Tailwind CSS 3, shadcn/ui, Radix UI
State	TanStack React Query
Routing	React Router v6
Backend	Supabase (PostgreSQL, Auth, Edge Functions)
AI	Google Gemini via Supabase Edge Functions
Charts	SVG progress rings, Recharts

---

📐 Architecture

src/
├── components/       # Reusable UI components (shadcn/ui based)
├── hooks/            # Custom hooks (useAuth, useMobile)
├── integrations/     # Supabase client & auto-generated types
├── pages/
│   ├── Auth.tsx       # Login / Signup
│   ├── Dashboard.tsx  # Metrics, progress rings, voice summary
│   ├── AddHabit.tsx   # Habit creation form
│   ├── HabitLogs.tsx  # Historical log viewer
│   └── HabitChat.tsx  # AI habit coach
└── lib/              # Utilities

supabase/
├── functions/
│   ├── habit-chat/    # AI coaching edge function
│   └── habit-summary/ # Weekly summary edge function
└── migrations/        # Database schema & RLS policies

---

🔒 Security

• Row Level Security (RLS) on all tables — users can only access their own data
• Auth-gated routes — all app pages wrapped in ProtectedRoute
• Server-side AI — habit data is processed in edge functions, never exposed client-side

---

🚀 Getting Started

Prerequisites

• Node.js 18+
• A Supabase project (or use the hosted version)

Setup

# Clone the repo
git clone https://github.com/<your-username>/habit-bloom.git
cd habit-bloom

# Install dependencies
npm install

# Create .env with your Supabase credentials
cp .env.example .env
# Fill in VITE_SUPABASE_URL and VITE_SUPABASE_PUBLISHABLE_KEY

# Start dev server
npm run dev

Environment Variables

Variable	Description
VITE_SUPABASE_URL	Your Supabase project URL
VITE_SUPABASE_PUBLISHABLE_KEY	Supabase anon/public key

---

📸 Key Screens

Dashboard

Real-time progress rings, streak tracking, weekly completion metrics, and one-tap voice summary.

AI Habit Coach

Personalized, data-driven coaching powered by Google Gemini — ask questions about your patterns and get actionable advice.

Add Habit

Create custom habits with emoji icons, color coding, frequency settings, and weekly targets.

---

🧠 What I Learned

• Designing secure multi-tenant data with Postgres RLS policies
• Integrating LLMs via edge functions with structured prompts grounded in user data
• Building accessible, responsive layouts with Tailwind + shadcn/ui
• Managing async server state with TanStack Query (mutations, invalidation, optimistic updates)

---

📄 License

MIT