Everything GMI does,
and how each piece works.

A guided first report.

A new account doesn't open on an empty page asking for a ticker. It asks you to pick a company you know — Apple, Amazon, Costco, Microsoft, Nike, Coca-Cola, Disney or Starbucks — and builds that report in one tap. The first time you open it, a few short notes point out what matters: this is the grade, here's what it's based on, here's what a low mark on debt means, and here's how to add your own holdings. Dismiss them and they're gone.

Not signed up yet? The two-minute walkthrough takes one real grade (Apple) apart factor by factor, before you create an account.

One grade scale, with a fair rule for banks.

Every stock is graded on the same published scale, so an A means the same thing for a chipmaker and a grocer. The one exception is written down: for banks, insurers and other Financial Services companies, debt-to-equity and current ratio are scored neutral, because leverage is how those businesses work, not a warning sign. Every other factor is unchanged, and the hover breakdown on the grade says so in place.

Beside the grade sit two style lenses, each a 0–100 reading with its own breakdown. The value lens reads free-cash-flow yield, P/E against a 15× baseline, dividend yield and price-to-book. The growth lens reads revenue, net income and free-cash-flow growth. A slow-growing cash generator can read "C+ grade, high value lens" — the letter stays comparable, and the lens adds the style context. How the grade works

Earnings Setup: the run-up, measured.

A company reports a solid beat and the stock falls anyway. Often the explanation is the run-up: by the time the numbers land, a lot of good news is already in the price. Earnings Setup puts that on one page for any company in the Earnings Lab calendar.

• The run-up chart — four years of prices with every earnings date marked by its result, and the move since the last report ranked against the company's own history (for example, "4th-largest run-up of the last 16 intervals").
• Sixteen quarters at a glance — EPS and revenue against estimates, how often the company beats, and a reaction table showing what the stock did the next session after each report, with the run-up into each one alongside.
• Where the bar is now — consensus estimates for the next report against the year-ago actuals, and whether estimates have been revised up or down.
• GMI's Read — a statistical read of the company's own history: how often the next session closed higher in setups like this one, a typical move range, the sample size, a confidence level, and a walk-forward record of how this read has held up in the past.

The fine print: GMI's Read is arithmetic on history, not a prediction and not advice. With at most sixteen reports per company, the probability is deliberately cautious; the move range is the steadier number. The recipe is on the page, with the numbers plugged in.

A stock screener over graded companies.

Filter the 2,400+ companies GMI grades by grade score, grade trend, sector, analyst sentiment, P/E, dividend yield, revenue growth, debt-to-equity, return on equity, free-cash-flow yield, operating margin and ML health. Start from a preset — Low P/E Large Caps, 3%+ Yield, Moderate Debt, Low Debt, Grade B or Better, Tech, 20%+ Revenue Growth or Improving Grades — or build your own and save it. Send the results to a symbol list or straight into a report.

Free tier: the top 25 matches of any screen and one saved screen. Paid accounts see every match and can save up to 50.

Investor Legends — famous criteria, test by test.

Eight screens built from criteria well-known investors and researchers published, each with its source cited on screen:

• Inspired by Peter Lynch — One Up on Wall Street
• Inspired by Benjamin Graham's defensive investor — The Intelligent Investor
• Inspired by Warren Buffett — quality criteria from Berkshire Hathaway annual reports
• Inspired by Joel Greenblatt — the Magic Formula
• Inspired by Joseph Piotroski — the F-Score
• Inspired by John Neff — John Neff on Investing
• Inspired by David Dreman — Contrarian Investment Strategies
• Inspired by William O'Neil — CAN SLIM

The part worth your time is the per-criterion breakdown. Every company gets a pass/fail line for each test — "passes 6 of 7 Graham tests; fails dividend record: 14 years, needs 20" — and a near-miss view shows companies one test away. Criteria the data can't support (like CAN SLIM's market-direction test) are marked "not evaluated", never guessed.

Save any legend or custom screen as a Smart List: a symbol list that refreshes on a daily, weekly or monthly schedule and logs every company that enters or leaves, with the reason ("left: PEG rose to 1.4, needs below 1"). On the free tier, a Smart List keeps the top matches within your eight tracked stocks.

The fine print: each screen is GMI's interpretation of published criteria. It is not affiliated with, endorsed by or reviewed by the investor named, and a company passing a screen is a description of its numbers, not a recommendation.

The Watchlist Dashboard.

Every symbol on your lists in one sortable grid: price, day change, volume, where it sits in its 52-week range, RSI, analyst sentiment, health score, grade and a 30-day sparkline. Filter by list, sector, sentiment or health, set it to refresh every 1, 5 or 15 minutes, and click any row for the detail.

Alerts — off until you turn them on.

GMI can email or text you when something changes in your holdings, your lists or the economy. Every alert starts off. You choose which ones you want on the Alerts page.

• Free alerts (email): a grade change on a stock you track, a change in the macro regime or macro grade, and a weekly Sunday digest.
• Pay-as-you-go alerts: earnings reminders and results against estimates, price moves, 52-week highs and lows, 50/200-day average crosses, drawdowns from your recorded cost, analyst consensus changes, list and portfolio grade-average moves, economic release briefings, sector grade changes and commodity moves.
• Prices: $0.004 per email (a digest counts as one email) and $0.06 per text message segment. You set a monthly cap — $5 unless you choose another — and it is never exceeded. At the cap, paid alerts pause until the 1st; free alerts keep coming.

Paid alerts need a card on file and a consent screen that shows the price list and a live monthly estimate from your own lists. Every message and its price is listed in your delivery history. And when GMI changes its own grading formula, you get one announcement — not a flood of grade-change alerts about companies where nothing happened. Alert pricing

Recurring reports that explain what changed.

Put a report on a schedule — weekly, every two weeks, monthly or quarterly — for any of your symbol lists. After each run, Claude compares the new report with the previous one and writes a change analysis: what improved, what weakened, and what's new. You get one notification when both are ready, and a Report History timeline keeps every run.

Cost: the usual $0.30 per stock for each report plus the Claude analysis, both estimated per run and per month before you save the schedule.

Portfolio benchmarking.

Compare a portfolio with a benchmark ETF — SPY, QQQ, ONEQ, DIA, IWM, VTI or any ticker you choose — over 1 month to 5 years. You see both returns side by side, the difference in points, a growth curve from the same starting dollars, risk measures (Sharpe ratio, beta, correlation, tracking error, maximum drawdown and more), and how much each holding contributed. The comparison is also written into the report's Excel workbook.

The fine print: returns are price-only (dividends aren't counted yet), and because portfolios don't carry purchase dates, today's holdings are measured as if held through each window. Holdings without enough price history are listed, not silently dropped.

A report card for our own forecasts.

The Macro Dashboard publishes machine-learning forecasts for economic series and commodities, each with an 80% range. The forecast report card grades those forecasts as they come due: how often the actual number landed inside the range, whether misses ran high or low, how often the direction was right, and whether the forecast did better than simply assuming "no change".

Every forecast is recorded when it's published, the first version counts, and nothing is scored until the month it targeted has ended and the data is out. Forecasts that can't be scored are counted and disclosed. Tour the Macro Dashboard