ORION-BI - Multi-Tenant AI-Powered Business Intelligence SaaS

ORION-BI transforms unstructured business data (emails, calendars, PDFs, spreadsheets) into explainable insights for SMBs and agencies.

🏗️ Architecture

• Frontend: React (Vite) + TailwindCSS + Clerk Auth
• Backend: Node.js + Express + Clerk
• AI/Analytics: Python 3.11 (scikit-learn, sentence-transformers, spaCy, Ollama)
• Database: PostgreSQL (Supabase Free Tier)
• Storage: Supabase Storage
• Deployment: Vercel (Frontend), Render (Backend), GitHub Actions (Python Jobs)

🚀 Quick Start

Prerequisites

• Node.js 18+
• Python 3.11+
• PostgreSQL (or Supabase account)
• Clerk account (for auth)
• Ollama (for local LLM) or API key for cloud LLM

Setup

1. Clone and install dependencies:

# Backend
cd backend
npm install

# Frontend
cd ../frontend
npm install

# Python AI
cd ../python-ai
pip install -r requirements.txt

2. Database Setup:

# Run migrations
cd database
psql -U postgres -d orion_bi < schema.sql

3. Environment Variables:

Create .env files in backend/ and frontend/:

backend/.env:

DATABASE_URL=postgresql://user:pass@host:5432/orion_bi
CLERK_SECRET_KEY=sk_test_...
SUPABASE_URL=https://your-project.supabase.co
SUPABASE_KEY=your-supabase-key
JWT_SECRET=your-jwt-secret
PORT=3001

frontend/.env:

VITE_CLERK_PUBLISHABLE_KEY=pk_test_...
VITE_API_URL=http://localhost:3001

4. Start Development:

# Terminal 1: Backend
cd backend
npm run dev

# Terminal 2: Frontend
cd frontend
npm run dev

# Terminal 3: Python AI (if running locally)
cd python-ai
python main.py

📁 Project Structure

ORION-BI/
├── frontend/          # React + Vite app
├── backend/           # Node.js API
├── python-ai/         # ML/AI analysis modules
├── database/          # SQL schemas and migrations
├── scripts/           # Utility scripts
└── docs/              # Documentation

🔐 Multi-Tenancy

All data is isolated by tenant_id (Clerk org_id). Row-level security enforced at database and API levels.

📊 Data Ingestion

• Email: IMAP integration or forwarding
• Calendar: Google Calendar API or .ics upload
• Documents: PDF, CSV, XLSX via upload
• Manual: Forms for ground truth data

🤖 AI Features

• Pattern analysis (seasonality, trends)
• Anomaly detection
• Forecasting (ARIMA, exponential smoothing)
• Natural language question answering
• Explainable insights with confidence scores

🚢 Deployment

See docs/DEPLOYMENT.md for detailed deployment instructions.

📝 License

MIT