Findian: AI-Powered Market Assistant 📈🤖

A conversational fintech platform that allows users to interact with the stock market using natural language via Telegram Bot. Findian bridges professional-grade data with a simple, chat-based interface.

🚀 Current Status: 85% Complete

✅ Completed Features (MVP)

1. Core Architecture

   • ✅ Microservices: FastAPI Backend + Telegram Bot Frontend
   • ✅ Database: MySQL (Relational) + Redis (Real-time PubSub)
   • ✅ AI Engine: Chutes AI (Llama 3.3) for Intent Parsing & Smart Clarification

2. Market Data Engine

   • ✅ Robust Feed: Pure Yahoo Finance (yfinance) integration
   • ✅ Hyper-Precision: 1-minute Price + Daily Volume/High/Low accuracy
   • ✅ Resilience: Mock Data simulation during API blackouts

3. Telegram Interface

   • ✅ Conversational UI: Natural language commands ("Price of Reliance", "Alert if TCS > 3000")
   • ✅ Financial Formatting: Consistent ₹1,234.50 format
   • ✅ Interactive Menus: Screener, Portfolio, Disclaimer inline buttons

4. Core Features

   • ✅ Smart Alerts: Redis-backed instant monitoring
   • ✅ AI Screener: Custom scans ("RSI < 30") + Pre-built strategies
   • ✅ Basic Portfolio: Add/View holdings with real-time P&L

⏳ In Progress (Next 15%)

1. Portfolio Management (Critical)

   • ⏳ Sell Logic & Position Exit
   • ⏳ Tax & Fees Calculation
   • ⏳ Daily Portfolio Reports (5:00 PM cron)

2. Monetization & Security

   • ⏳ User Tiers (Free vs Premium limits)
   • ⏳ Access Control for Pro features
   • ⏳ API Key Encryption

3. Cloud Deployment

   • ⏳ Webhooks for <100ms latency
   • ⏳ Production Server (AWS/DigitalOcean)

Core Capabilities

⚡ Data Engine

• Primary: Yahoo Finance (yfinance) for 100% uptime and 1-minute precision
• Stability: Zero-login architecture with no session management required

🔔 Intelligent Alerts

• Redis-backed monitoring system
• Real-time price target and technical indicator tracking
• Instant notifications to users

🔍 Auto-Screener

• Continuous Nifty 50 scanning
• Technical signals (RSI, breakouts, volume spikes)
• Custom AI Scans based on natural language

🛠️ Technology Stack

Component	Technology
Frontend	Python-Telegram-Bot
Backend	FastAPI
AI Engine	Chutes AI (Llama 3.3)
Database	MySQL + Redis
Data Feeds	Yahoo Finance (yfinance)
Container	Docker
Memory System	Agent Memory for AI persistence

🚀 Quick Start

1. Clone the repository

   git clone http://localhost:8443/sanprat/Findian.git
   cd Findian

2. Set up environment

   cp .env.example .env
   # Edit .env with your API keys

3. Run with Docker

   docker-compose up -d

4. Start the bot

   python bot/main.py

5. Try commands on Telegram

   • /start - Begin your journey
   • "Price of Reliance" - Get current price
   • "Alert me when TCS crosses 3500" - Set price alert
   • "Show me stocks with RSI < 30" - AI screener

🔐 Security Checks

Before committing or pushing, run the repo secret scan:

pre-commit run --all-files

.env must remain gitignored and never be committed.

📋 Agent Memory System

Findian includes an innovative Agent Memory System that enables AI agents to:

• Learn from user interactions
• Remember successful strategies
• Share knowledge across sessions
• Maintain project continuity

All AI agents must follow memory guidelines in GEMINI.md

📁 Project Structure

Findian/
├── backend/                    # FastAPI backend services
│   ├── app/
│   │   ├── core/              # Core trading logic
│   │   │   ├── ai.py          # AI and NLP processing
│   │   │   ├── market_data.py # Data integration
│   │   │   ├── scanner.py     # Auto-screener logic
│   │   │   └── scheduler.py   # Task scheduler
│   │   ├── db/                # Database models
│   │   └── main.py            # FastAPI entry
├── bot/                       # Telegram bot frontend
│   └── main.py                # Bot logic
├── agent_memory.sh            # Memory system
├── universal_agent_hooks.py   # Multi-agent integration
└── docker-compose.yml         # Container orchestration

🔮 Roadmap

• Portfolio sell functionality
• Tax calculation engine
• Daily portfolio reports
• Premium subscription tiers
• Cloud webhook deployment
• Mobile app companion

📊 Usage Examples

Natural Language Commands

User: "What's the price of Infosys?"
Bot: "₹1,456.30 ▲ +12.50 (+0.86%)"

User: "Alert me if Reliance goes above 2500"
Bot: "✅ Alert set: Notify when RELIANCE > ₹2,500"

User: "Show me oversold stocks"
Bot: "🔍 Found 5 stocks with RSI < 30: [List]"

🤝 Contributing

This project uses AI agent collaboration with memory persistence. All agents follow guidelines in GEMINI.md for consistent development.

---

Findian: Democratizing algorithmic trading through conversation 💬📈

📜 License

This project is licensed under the GNU Affero General Public License v3.0 - see the LICENSE file for details.