It watches the S&P 500 on daily closes and tells you when something crosses a level you chose — and, before you agree to it, how often that alert has fired and how often the crossing didn't hold.
Every figure in this deck was read from the live API at the close of 14 September 2026, and is stamped rather than refreshed — the same rule the app applies to itself.
Short-horizon, about the last fortnight of trading, and measured against this stock's own recent range. It says nothing about whether the company is any good.
Earnings, price strength, quality, buying pressure and sector, ranked across the index. It says nothing about whether today is a sensible day to act.
Every stock has both, always — and one page shows them together, with whichever one you arrived through leading and the other a tap away.
Neither is a recommendation. They are two measurements, published side by side, and the app never combines them into a verdict.
RSI measures how hard a stock has been bought or sold over the last 14 sessions, on a 0–100 scale. Most of the time it sits in the middle and means nothing. Its value is how rarely it speaks.
Daily closes, not hourly. An hourly chart crosses these levels constantly and most of it is noise. A stock reaching oversold is a roughly twice-a-year event, which is what makes it worth a notification.
"RSI 63" tells you very little. What matters is where that sits in this stock's own recent range, and how much this stock moves in a day. Both travel with every reading.
Why it matters. A five-point RSI move on a stock that travels 0.8% a day is a different event from the same move on one that travels 4%. That day 241 of 503 were ranging and 145 trending — the same reading means different things in each.
Not a score out of 100 with a fixed pass mark. A percentile: 90 means top 10% today. So it moves when the field moves, and it cannot drift into everything looking good at once.
47 of 503 was 9.3% — close to a tenth, because that is what a percentile rank means rather than a threshold someone picked.
Tap any score and it opens into its parts, with the weights the scoring job actually ran with — read back from the job, not typed into the app.
And it tells you it isn't ready. Ask the app how deep this history goes and it answers: 13 days, "too little to read a frequency from — treat any number as provisional." That sentence is in the product, not in a disclaimer.
A price pokes through your level during the day and closes back under it more often than not. Alerts are judged on the closing price, so that poke is filtered out before it ever reaches your phone.
How this was measured. 500 sessions to 2026-09-14, eight widely held S&P 500 stocks, at round levels across each stock's range.
A stock that stays oversold for three weeks is one notification, not twenty-one.
Every alert shows how often it would have fired on the stocks you picked — measured on the same bars that will decide it.
A stock with too little history is named, not quietly dropped. A score that couldn't be read says so rather than showing zero.
One notification per weekday, naming everything that crossed, evaluated after the close on the settled daily bar — not on a price that is still moving.
Create an account, know your favorite companies and set meaningful alerts, so you can find out when the condition is right.
Currently in alpha on TestFlight, US-listed S&P 500 constituents only. JujsyAlpha is not a broker and does not place orders.