MarketMentor

MarketMentor teaches complete beginners how the stock market works, using live data, an autonomous research agent, and adaptive quizzes.

Comment

MarketMentor: The AI Tutor That Turns Complete Beginners Into Confident Investors

Inspiration

Over 50% of Americans have no stock market exposure. Not because they lack money, but because they lack understanding. The financial world speaks in jargon: P/E ratios, EPS, volatility, moving averages. For someone who grew up without financial education, opening a brokerage account feels like reading a foreign language.

We built MarketMentor because financial literacy should not be a privilege. It should be a right. Every person, regardless of background, income, or education, deserves to understand how markets work and how to grow their wealth. MarketMentor is the AI tutor that makes that possible.

What It Does

MarketMentor is a full-stack AI-powered financial literacy platform with four core modules:

Learn: An AI tutor powered by Claude that explains any financial or economic concept in clear, accessible language. From "What is a stock?" to "How do interest rates affect bond prices?" Users can tap a Quick Topic or ask anything in natural language. Every explanation includes a real-world analogy, a breakdown, and a key takeaway.

Market Research Agent: An autonomous AI agent that researches any stock on demand. Users type a question like "Should a beginner invest in NVIDIA?" and the agent streams its thought process live: fetching real market data, analyzing the 52-week range and P/E ratio, assessing volatility and risk level, then generating a beginner-friendly investment brief. Powered by live Yahoo Finance data with no stale information.

Markets: A real-time dashboard showing live indices (S&P 500, NASDAQ, Dow Jones, VIX) and individual stocks. One click generates an AI summary of what is moving the market today and why, turning raw numbers into actionable context.

Paper Trade: A risk-free virtual trading simulator with $10,000 in virtual funds. Users practice buying and selling real stocks at real prices, track their portfolio performance, and consult the AI Portfolio Advisor for personalized analysis of their holdings and diversification strategy.

Quiz: An adaptive quiz engine that generates fresh financial literacy questions at three difficulty levels. After each answer, the AI explains exactly why the correct answer is correct, reinforcing learning rather than just testing it.

How We Built It

Frontend: Next.js 14 with App Router, chosen for its built-in API routing, server-side rendering, and seamless Vercel deployment. Clean, minimal UI with Inter font, a white/black/gray palette, and color-coded financial indicators (green for gains, red for losses).

Backend: Next.js API routes handle all server-side logic with no separate Express server needed. This keeps the codebase unified and the architecture clean.

AI: Claude API (claude-sonnet) powers every intelligent feature including the concept explainer, the Market Research Agent, the market summarizer, the portfolio advisor, and the quiz generator. Each feature uses a carefully engineered system prompt to ensure responses are always beginner-appropriate, accurate, and concise.

Market Data: yahoo-finance2 npm package pulls live stock prices, historical data, 52-week ranges, P/E ratios, and index values in real time. MIT licensed. No API key required.

Streaming: The Market Research Agent uses Server-Sent Events (SSE) to stream its step-by-step thought process to the frontend in real time, giving users full transparency into how the AI reaches its conclusions.

Challenges We Ran Into

The hardest part was designing the Market Research Agent's streaming pipeline. Getting SSE to work reliably with Next.js App Router required custom response handling that does not follow the standard fetch pattern. After several failed approaches, we settled on a ReadableStream-based implementation that flushes each agent step as it completes.

The second challenge was prompt engineering. Getting Claude to explain financial concepts at exactly the right level, detailed enough to be educational but simple enough for a true beginner, required dozens of iterations. Too technical and users feel lost. Too simple and experienced users feel talked down to. We landed on a four-part structure: definition, analogy, breakdown, key takeaway.

Accomplishments We Are Proud Of

  • A fully functional agentic AI that fetches real data, reasons about it, and explains it (not a mock or simulation)
  • Live market data integration that makes every session feel current and relevant
  • A UI clean enough that a 15-year-old and a 50-year-old can both navigate it without confusion
  • A quiz engine that generates genuinely different questions every time with no hardcoded question bank

What We Learned

Building MarketMentor taught us how much the framing of financial information matters. The same P/E ratio data lands completely differently depending on whether you say "the stock trades at 68x earnings" versus "investors are paying $68 for every $1 of annual profit." People do not struggle with the numbers. They struggle with the meaning. That insight shaped every design decision in this app.

What's Next

  • Watchlist and Alerts: Users save stocks and get AI-generated updates when significant moves happen
  • Learning Paths: Structured 7-day courses covering Investing 101, Understanding Bonds, and How to Read an Earnings Report
  • Social Portfolio: Anonymized leaderboard where users compare paper trading performance and learn from top performers
  • Mobile App: React Native port for on-the-go financial learning
  • Multilingual Support: Expanding to Spanish, Filipino, Hindi, and Mandarin to serve underrepresented communities globally

Built for Hackonomics 2026. Fusing computer science and economics to make financial literacy accessible to everyone.

Built With

Share this project:

Updates