fundamental-agents 2025

projects

View on GitHub

fundamental-agents demo
An orchestrator runs five agents in sequence — data gathering, financial metrics, news sentiment, valuation, synthesis — turning a ticker into an investment report.

A multi-agent AI platform for fundamental stock analysis. Input a ticker and get an institutional-style investment report produced by specialized agents that gather data, compute metrics, analyze sentiment, and model valuations, coordinated by an orchestrator.

The agent pipeline

The orchestrator runs five agents in sequence (shown in the animation above):

  • Data Gathering fetches financial statements, price history, company profile, and recent news (Financial Modeling Prep, NewsAPI).
  • Financial Metrics computes P/E, D/E, ROE, and the current ratio from the raw financials.
  • News Sentiment VADER sentiment analysis over recent articles.
  • Valuation a discounted-cash-flow model (WACC, projected free cash flow, terminal value).
  • Synthesis assembles every agents output into a markdown report with a recommendation.

Full-stack app

It is a complete product, not just a script:

  • Frontend Next.js 13 (App Router), React, TypeScript, Tailwind, with auth, a dashboard, a report viewer, and pricing pages.
  • Backend FastAPI + SQLAlchemy + Pydantic, PostgreSQL, JWT auth, and Stripe checkout/webhooks.
  • Infrastructure Docker Compose, Caddy reverse proxy, and GitHub Actions.