AI Investment Research Agent

AI-powered investment research platform — search any public company and get an Invest/Pass decision with financial data and news analysis.

A production-ready AI investment research agent built with a Next.js 14 frontend and FastAPI backend. The system accepts a company name, gathers financial data via yfinance and news via DuckDuckGo, then uses an LLM (OpenAI GPT-3.5-turbo or Groq Llama 3) to return an Invest or Pass decision with detailed reasoning. Features JWT authentication, portfolio tracking, research history with pagination, and a neo-brutalist UI with dark mode support.

Tech Stack

Next.js 14FastAPITypeScriptPythonLangChainyfinanceDuckDuckGo SearchSQLiteJWT

Features

AI-powered Invest/Pass decision for any public company
Real-time financial data via yfinance (price, P/E, market cap, revenue growth)
News and web data via DuckDuckGo search
LLM-powered reasoning with OpenAI GPT-3.5-turbo or Groq Llama 3
JWT authentication with signup/login flows
Portfolio tracking — create portfolios, add/remove holdings
Paginated research history with search and deletion
Neo-brutalist UI with dark mode, Framer Motion animations

Challenges

Designing a linear AI pipeline that gracefully handles partial tool failures.
Implementing unified API abstraction across OpenAI and Groq providers.
Building a responsive neo-brutalist UI with smooth Framer Motion animations.
Securing JWT auth with proper password hashing and token expiration.

Learnings

Built an AI agent pipeline with tool orchestration and JSON validation.
Implemented full-stack auth with JWT and SQLite storage.
Designed a neo-brutalist design system with dark mode support.
Integrated multiple LLM providers with fallback support.

Feedback

For feedback or suggestions, contact me at: shouriyatayal1234@gmail.com

Links

Live Github