Quality scores, reverse-DCF valuation, filings and a portfolio record.
StockIQ scores a company against a fixed set of tests, values it from its own reported cash flows, and shows what institutions, members of the US House and company insiders have filed. Every input is public: company reports, market data and regulatory filings.
The 7-day trial opens everything on Desk. No card required. It falls back to Sandbox when it ends, rather than locking you out.
What the system holds
Listings by market: United States 6,727 · London 368 · Tokyo 225 · Frankfurt 39. One further listing sits on a fifth exchange and is not counted as a covered market, so the four figures above are one short of the tracked total.
Measured against the live database on 26 August 2026. The counts move as the nightly batch runs.
8-Pillar quality score
A count, not an opinion.
Each pillar is one test against a reported figure, and each test passes or fails. Nothing is weighted out of sight and nothing is smoothed into a single unexplained number.
Every failure shows the figure that caused it.
A company that scores six shows the two tests it failed and the number that failed them, so the result can be checked against the filing.
Percentile ranks and score history.
Quality and valuation ranks put a score in the universe rather than only on a scale. Score history shows when a company started failing a test it used to pass.
Illustrative recreation of the pillars panel. The figures are examples; the eight tests are the ones the model runs. The threshold each test is measured against is shown next to it in the app.
Illustrative recreation of the decision-signal row. Each chip is one computed value: green where the measurement reads well, amber where it does not, plain where it is neutral. There is no overall verdict chip, because a rule-based tool cannot know your circumstances.
Reverse-DCF fair value and the margin-of-safety grid
A single fair value is false precision. What matters is which assumptions have to be true for today's price to make sense, and how fast the answer moves when they are not.
| Assumption | Where its starting value comes from |
|---|---|
| Revenue growth | Derived from the reported revenue history. |
| Net margin | Derived from the trailing reported margin. |
| Exit P/E | Derived from the company's own trailing multiple. |
| Years in the projection | A fixed default of five years — the projection the diagram above runs. Change it. |
| Required return | A fixed default. Change it. |
The arithmetic runs in the browser. Moving a slider does not call the server, so the grid redraws as you drag rather than after a round trip.
| Exit P/E | Required 8% | Required 9% | Required 10% | Required 11% | Required 12% |
|---|---|---|---|---|---|
| 14x | -10%£38 | -12%£37 | -17%£35 | -21%£33 | -24%£32 |
| 18x | +17%£49 | +12%£47 | +7%£45 | +2%£43 | -2%£41 |
| 22x | +43%£60 | +38%£58 | +31%£55 | +26%£53 | +19%£50 |
| 26x | +69%£71 | +62%£68 | +55%£65 | +48%£62 | +40%£59 |
Illustrative recreation of the convergence grid. Example figures. Rows are the exit P/E. Columns are the annual return you require of the investment — the rate the future value is discounted back at, not a return the company produces. Each cell is the margin of safety at that pair of assumptions; the smaller figure is the implied fair value, and the outlined cell is the current assumption set. Reading across a row shows how much of the answer is the return you demand; reading down a column shows how much of it is the multiple you assume at the end.
13F, Congressional and Form 4 filings
All three regimes are read from the primary source — the filing itself — and every position carries the date it was filed.
What a 13F does not contain
Filed by institutional managers above the SEC reporting threshold, a 13F is a snapshot of long US-listed equity positions on the last day of a quarter. It omits shorts, cash, bonds and non-US listings, so it cannot be turned into a fund's return. It is a list of companies worth looking at, not a scoreboard.
Purchases, not every Form 4 event
Form 4 is filed by company directors, officers and holders of more than 10 per cent. The census counts open-market purchases: grants, option exercises and scheduled disposals are all Form 4 traffic and none of them means what an officer buying at the market price with their own money means. One filing can cover several reporting owners, so the count is by person, not by document.
The model screens
The same universe is ranked by each screen in turn. A company can be excellent on one and unremarkable on the others, and that disagreement is the useful part: it says what kind of case there is rather than collapsing separate questions into one number.
Illustrative recreation of the model-score panel. Example figures. Amber marks a rank in the bottom half of the universe for that model.
| Model | What it ranks on | What it excludes, and why |
|---|---|---|
| Quality | How many pillar tests a company passes, and by how far it clears each one. | No sector exclusions. |
| Deep Value | Enterprise-value cheapness. What you pay for the operating business, against what it earns and what it returns on capital. | Financials and Real Estate excluded outright: for a bank, debt is raw material rather than a liability to net off, so an enterprise-value multiple is a category error. A Piotroski F-Score of 2 or below is also excluded, but only where there is enough reported data to score it; where there is not, the company is carried as unknown. |
| Cheap + Catalyst | A cheap valuation rank combined with a recent change on the company, so the list is not simply the names that have been cheap for years. | No sector exclusions. |
| Smart Money | How many tracked institutional filers hold the company, and whether the most recent filings added to the position or cut it. | Nothing is inferred beyond what was filed. |
Holdings, lots and portfolio P/L
Nothing is fetched from an institution.
No Open Banking consent, no brokerage API key, no OAuth grant and no aggregator anywhere in the product. Holdings sit in accounts you create and name, so a name you type is a label for your own benefit and authenticates nothing.
The caps are on what you add, not on what you already hold.
Sandbox tracks 3 holdings across 2 labels, with up to 5 buy lots on each, and Analyst removes the caps. If a trial ends while you are tracking more than Sandbox allows, nothing is deleted and the holdings stay readable; you simply cannot add another until you upgrade.
- Buy lots carry a date, a share count, a price and the fees you paid, so cost basis is per lot rather than an average you cannot reconstruct.
- Disposals are recorded separately, so a sale leaves a record instead of disappearing from the position.
Lots are typed in. There is no broker CSV importer for them.
- Unrealised and realised profit and loss, reported in GBP.
- US positions and London pence prices are both converted to GBP, so a mixed book adds up rather than adding dollars to pence.
- Benchmark reports against VOO and QQQ, so the question is whether the book beat an index fund.
- Platform-fee and fund-cost analysis across your labels, including the ongoing charges inside funds.
- A monthly position review that re-runs every holding through the same rule set and classifies it. It is the model's output on the companies you own, computed identically for every user from public data, with no knowledge of your circumstances. It is not advice.
A record of what you did and why: the thesis at entry, what you expected, what happened, and what you got wrong. CSV import exists here, for loading existing history in bulk. It does not populate the lot tracker.
Telegram alerts on six conditions
Six conditions, and the evaluator recognises no seventh.
Every rule is written against a value the system already computes.
A fifteen-minute cycle, with a cooldown on every rule.
Rules run through the trading day. The cooldown is what stops a price sitting on your level from sending the same message forty times before lunch.
An alert reports that a condition you set has been met. It does not tell you to do anything about it.
Six things it does not do
- No broker or bank connection. No Open Banking, no brokerage API key, no OAuth link, no read-only account connection. There is nothing to authorise and nothing to revoke.
- No trade execution. StockIQ cannot place an order, cancel one, or size one. It has no route to a market.
- No real-time tick data. The unit of analysis is the daily bar. Watchlist quotes refresh on a few-minute cycle while markets are open, delayed rather than streaming. This is not a trading terminal.
- No lot import. Buy lots are typed in by hand. The CSV import belongs to the trading journal.
- No personal advice. Every output is rule-based and computed the same way for every user from public data. It knows nothing about your circumstances, tax position or objectives.
- No two-factor authentication. There is no authenticator-app or SMS second factor in the product today. It is on the roadmap. Until it ships the account is protected by a password alone, so use one you use nowhere else.