Quant Flow AI

"AI-powered stock scanner that turns plain English into institutional-grade analysis — no coding required."

Comment

URL

https://quantflow.bjjwwangs.win/

Inspiration

As retail investors, we often struggle to keep up with "NO CODE" stock selection tools. We wanted to create an accessible yet powerful system that combines traditional quantitative analysis with modern AI capabilities, making professional-level stock screening available to everyone.

What it does

QuantFlow is an intelligent stock scanning and analysis platform that helps investors discover opportunities using:

  • 5-Phase Classification: Automatically categorizes stocks into phases (bottom, inflection, moderate uptrend, strong uptrend, weekly-strong-daily-flat)
  • Customized Strategy: User can input natural language to define the strategy, with which Gemini 3 Pro generates Python code to analyse the whole market.
  • Technical Indicators: MACD, Moving Averages, Bollinger Bands, Hurst Bands, Volume Analysis
  • Fundamental Screening: Filter by sector, market cap, institutional holdings, and insider activity
  • AI-Powered Analysis: Custom natural language filters via Gemini AI, deep audit analysis, and intelligent chatbot

How we built it

  • Backend: Python with FastAPI, using yfinance for market data and SQLite for storage. Use Antigravity and Vibe-Kanban(Gemini and Claude).
  • Frontend: React + TypeScript with Tailwind CSS for a clean, responsive UI. Mainly use AIStudio and Antigravity, refined by vibe-kanban(Gemini)
  • AI Integration: Google Gemini API for natural language strategy translation and stock analysis (Gemini-3-Pro and Flash)
  • Technical Stack: Custom implementations of quantitative indicators and phase classification algorithms

Challenges we ran into

  • Accurately estimating institutional cost basis using VWAP calculations over 60-day windows
  • Safely executing AI-generated Python code in a sandboxed environment for custom filters
  • Balancing data freshness with API rate limits from Yahoo Finance
  • Creating an intuitive UI that doesn't overwhelm users while exposing powerful features

Accomplishments that we're proud of

  • Built a complete end-to-end stock analysis pipeline from raw data to actionable insights
  • Implemented a custom AI filter where users describe strategies in plain English
  • Achieved institutional-grade phase classification that identifies market opportunities
  • Created a responsive, professional UI with real-time K-line charting

What we learned

  • The complexity behind professional trading platforms and quantitative analysis
  • How to integrate multiple data sources reliably for financial applications
  • The importance of sandboxing when executing AI-generated code
  • Balancing feature richness with user experience simplicity

What's next for QuantFlow

  • Real-time streaming quotes and alerts
  • Portfolio tracking and position management
  • Backtesting engine for strategy validation
  • Mobile app for on-the-go monitoring
  • Multi-market support (A-shares, crypto)

Built With

Share this project:

Updates