Signal Forge is an AI-powered financial analyst that lives in your browser. Unlike simple chatbots, it uses LangGraph to orchestrate a multi-step research workflow. It autonomously plans, searches live news (Tavily), analyzes market data (yfinance), and generates institutional-grade investment memorandums with clear "Bullish/Bearish" verdicts.
|
Landing Page Clean, distraction-free entry point. 
|
Secure Authentication OAuth integration via Supabase Auth. 
|
|
The Agent at Work Real-time logging of reasoning steps. 
|
The Final Report Complete memo with charts & verdict. 
|
Dashboard (Mobile)
|
Settings (Mobile)
|
- 🧠 Agentic Workflow: Powered by LangGraph & LangChain to handle complex reasoning loops.
- 🚀 High-Performance Caching: Uses Redis (optional) for efficient state management and caching.
- 📊 Live Market Data: Fetches real-time price history and technicals using
yfinance. - 🌍 Real-Time News: Scrapes and synthesizes the latest financial news via Tavily API.
- 🔐 Secure Auth: Complete OAuth 2.0 implementation with Supabase Auth.
- 💾 Persistent History: Autosaves every report to Supabase (PostgreSQL) via SQLAlchemy.
- 🎨 Responsive UI: A "Soft Dark Mode" interface built with Tailwind CSS, optimized for both desktop and mobile.
| Frontend | Backend & AI | Data & Infra |
| Next.js 14 (App Router) | Python 3.11 (FastAPI) | Supabase (PostgreSQL) |
| TypeScript | LangChain & LangGraph | Supabase Auth |
| Tailwind CSS | Google Gemini 2.5 Flash | Redis (Caching) |
| Framer Motion | Tavily Search API | Vercel & Render |
Kalash Pratap Gaur
@CaSh007s
Built with ❤️ using Google's GenAI Stack.