An AI-powered self-discovery platform that helps users overcome Solomon's Paradox — the tendency to reason more wisely about others' problems than our own. Zen creates a Digital Twin from your journal entries using RAG and pairs it with a Council of 50+ historical mentors to provide personalized wisdom and guidance.

Built at UofT Hacks 2025.

Write and reflect on journal entries with AI-generated follow-up questions that deepen self-awareness. Entries are embedded and stored for semantic retrieval, allowing the system to learn about you over time.

Chat with a multi-agent system that intelligently routes your conversation through:

• The Empath — initial intake and emotional recognition
• Discovery Agent — clarifying questions to understand your situation
• Wise Mentor — personalized guidance from 50+ historical figures (Marcus Aurelius, Maya Angelou, Steve Jobs, Socrates, and more)

Mentors are auto-selected based on your situation and reference your journal history for context-aware advice.

Personalized meditation sessions (5–30 min) streamed in real-time with:

• 5-stage journey: welcome, breathing, body scan, visualization, closing
• Breathing circle animation and progress tracking
• Content personalized from your journal entries
• Post-meditation reflection capture

AI-generated analysis of your journal patterns and emotional landscape — your Digital Twin's perspective on who you are.

├── backend/
│   ├── app/
│   │   ├── agents/orchestrator.py    # LangGraph workflow + mentor personas
│   │   ├── routers/                  # journal, chat, meditation, digital_self
│   │   ├── services/                 # RAG engine, personalization
│   │   ├── models/schemas.py         # Pydantic models
│   │   ├── main.py                   # FastAPI entry point
│   │   ├── config.py                 # Settings
│   │   └── database.py               # Supabase client
│   ├── requirements.txt
│   └── supabase_schema.sql
├── frontend/
│   ├── app/                          # Pages: journal, counsel, meditation, digital-self
│   ├── components/zen/               # ParticleBackground, Navigation
│   └── lib/                          # Utilities

• Python 3.10+
• Node.js 18+
• Supabase account
• Google Gemini API key

1. Go to your Supabase project dashboard
2. Enable the pgvector extension
3. Run backend/supabase_schema.sql in the SQL Editor

cd backend
python -m venv venv
source venv/bin/activate  # Windows: venv\Scripts\activate
pip install -r requirements.txt
cp .env.example .env

Edit .env:

SUPABASE_URL=your_supabase_project_url
SUPABASE_KEY=your_supabase_anon_key
SUPABASE_SERVICE_KEY=your_supabase_service_role_key
GOOGLE_API_KEY=your_google_gemini_api_key

python -m app.main
# http://localhost:8000

cd frontend
npm install
cp .env.local.example .env.local

Edit .env.local:

NEXT_PUBLIC_SUPABASE_URL=your_supabase_project_url
NEXT_PUBLIC_SUPABASE_ANON_KEY=your_supabase_anon_key
NEXT_PUBLIC_API_URL=http://localhost:8000

npm run dev
# http://localhost:3000

User -> Next.js Frontend -> FastAPI Backend -> LangGraph Orchestrator
                                |                    |
                                v                    v
                           Supabase             Google Gemini
                        (pgvector RAG)        (embeddings + chat)

The LangGraph orchestrator manages a state-based workflow that routes conversations through specialized agents. The RAG pipeline embeds journal entries using Google's text-embedding-004 model and retrieves semantically relevant context via pgvector, injecting it into mentor responses for personalized guidance.

Interactive API docs available at http://localhost:8000/docs when the backend is running.

MIT