Personal project

An equity research tool I built for my own investing.

I wanted proper research on the stocks I hold without paying for a terminal, so I built it on free, official data: SEC filings, FRED and reported financials.

research.ryanappuhamy.com · Research report
Research report for AAPL: price, 1-year change, 52-week range and chart
Research report

Fact and interpretation, kept apart

Numbers from APIs and filings are stored as-is. The AI note is labelled, attributed, and instructed never to reference a figure that isn't in the input. Template output takes over if no model is reachable.

  • Six metric cards: valuation, growth, profitability, financial health, TTM financials
  • Insider transactions parsed from raw Form 4 XML: name, role, buy / sell, size, date
  • Earnings guidance pulled from the 8-K. The legal boilerplate is trimmed first, so the forward outlook survives the token budget
…/?ticker=AAPL · fundamentals and insider activity
Fundamentals cards and SEC Form 4 insider activity for AAPL
Under the hood

One request, a seven-step pipeline

Every ticker lookup runs the same ordered workflow. Each step fails gracefully (a missing source returns a note, never a broken report) and results are cached in Postgres, so a repeat lookup is free.

  1. 1Price history & stats
  2. 2Fundamentals
  3. 3Peer comparison
  4. 4SEC insider activity
  5. 5Transcript + AI read
  6. 6Analyst consensus & macro
  7. 7AI research note
SourceCostWhat it provides
SEC EDGARfreeForm 4 insider filings and 8-K earnings exhibits, the institutional layer most retail tools skip
yfinancefreeDaily prices, portfolio quotes, fallback fundamentals
Finnhubfree tierLive quotes, analyst trends, earnings calendar, company news and IP-independent fundamentals for when cloud hosts get blocked
FREDfreeFed funds rate, 10-year yield, CPI, unemployment
Gemini → Groq → Claudefree-firstResearch notes and weekly briefs. The chain only reaches a paid model if both free tiers fail
Portfolio

Your portfolio, read like a fintech app

Live prices and P&L, with the number that matters first and the detail one tap away.

  • Scrubbable value chart vs the S&P 500 and other indices
  • Portfolio health: return, risk, diversification and valuation, each in one plain sentence
  • Tax if you sold today, by country
  • Honest about the chart: it's a backtest of current holdings
…/portfolio
Portfolio: value, today's P&L, scrubbable chart vs the S&P 500 and holdings
…/portfolio · health
Portfolio health: return vs market, risk, diversification and valuation
Simulator

What a monthly investment plan really does

A dollar-cost-averaging simulator on 15 years of real monthly prices, for index ETFs or any stock you hold.

  • Historical mode vs a Monte Carlo projection (10th to 90th percentile)
  • Fees, inflation and capital-gains tax by country
…/simulator
Investment plan simulator on 15 years of real prices
Alerts

Metric rules, checked against the live pipeline

Price, P/E, revenue-growth or insider-filing thresholds. The list loads instantly; the condition check runs in the background and fills in the status in place.

  • Each alert reads as one line ("NVDA · Price above $230.00") with an OK or Triggered badge, so what needs attention is visible at a glance
…/alerts
Alerts with OK and Triggered status
Macro calendar

The week's data releases, in context

A dedicated view of upcoming US economic events: the releases the weekly brief keeps naming as near-term risks.

  • Actual vs. forecast for ISM, JOLTS, non-farm payrolls, CPI and more
…/macro
Macro calendar of upcoming US economic releases
Built with
Python · FastAPINext.js 16Supabase PostgresRenderVercel

An educational project, not financial advice. Every number is traceable to a primary source (SEC EDGAR, FRED, reported financials); analyst figures are passed through with attribution.