Fair questions, straight answers

No. RockTurners publishes impersonal, general-circulation research: the same methodology, criteria, and scoring applied identically to every company. It knows nothing about you, and its verdicts describe methodology fit — never an action to take. Consult a licensed financial professional before making investment decisions.

Action calls are advice, and advice depends on knowing the person receiving it. RockTurners publishes the same research to everyone, so its output stops at what the filings show: how well a company fits a published framework. What to do with that reading depends on circumstances only you — and a licensed professional — know.

Because that's the honest claim. A scan doesn't discover whether a company is good — it measures how well the filings match one framework's specific criteria. The same company can be a strong match to a dividend discipline and no match to a deep-value screen; both verdicts are correct, and neither is a statement about what happens next.

Gemstone: a strong match — the filings support the framework's criteria. Geode: a partial match — some criteria hold, others can't be verified from the filings. Sediment: no match — the criteria aren't met, kept as a turned-over stone for the record. All three describe framework fit — how well the company's filings satisfied the screening and evaluation criteria of the framework you selected. They describe the strength of that match. They are not a recommendation to buy, sell, or hold anything, and not a prediction of returns.

Yes — with one structural difference. Like any stock screener, RockTurners filters roughly 4,000 U.S.-listed companies with fixed quantitative criteria. Its second stage is what most screeners don't have: an AI evaluator that reads the survivors' SEC filings and returns a verdict on fit to a published framework, with every claim cited to its filing and every gap named as unknown. It screens and it reads — it never suggests an action.

Because they answer different questions. Most ratings elsewhere are timeliness judgments — price momentum, analyst estimate revisions, valuation against sector peers. A verdict here measures one thing: how well a company's SEC filings match a published methodology's criteria, as of the filings it read. The two can diverge legitimately: a quality-focused screen can surface a company precisely because its price has fallen, at the same moment a momentum model scores it poorly. Neither is broken — they measure different axes, which is also why rating sites so often disagree with each other. Every verdict names its filing dates; anything that happened after them is, by definition, not in it.

No — by design. The same filings produce the same output for every reader, and nothing about who you are is an input. Impersonal, general-circulation research is what keeps this publishing rather than advice.

Primary sources: SEC EDGAR filings — annual and quarterly reports, event disclosures, and insider forms — plus financial metrics computed from company facts. Every factual claim in a verdict cites the specific form, section, and date it came from, so you can read the source yourself.

A fixed discipline expressed in two stages: a quantitative screen that narrows the universe, and an evaluation rubric the AI applies to the survivors’ filings. Ten frameworks are built in, each in the tradition of published investing research and organized under the value, growth, quality, and income lenses. Alongside them, General Analysis is a general-purpose strategy for portfolio holdings and one-off company reads — it evaluates one named company rather than screening a universe, so it belongs to no lens and carries no scoring rubric.

Yes. The Strategy Lab lets you fork any built-in framework — forking copies it into your own editable version, so you can adjust its screening criteria (thresholds, metrics, universe scope) without changing the original. You can also describe a screen in plain language — metrics, thresholds, qualities, sectors — or paste a methodology document, and the Lab composes a screen you review before saving. What you describe are companies, never your own circumstances: the Lab does not take age, income, holdings, or risk tolerance as inputs. Derived frameworks run through the same pipeline with the same EDGAR grounding, citations, and honest unknowns as the built-ins.

That dimension is marked unknown — never inferred, never filled in from the model’s memory — and enough unknowns cap the verdict. An honest gap beats a confident fabrication; the unknowns list on each verdict shows exactly what could not be verified.

U.S. operating companies filing with the SEC on approved exchanges (Nasdaq, NYSE, NYSE American) — roughly 4,000 names after excluding funds, trusts, and blank-check entities. A framework’s own criteria can narrow that universe further.

Each verdict names the filings it read — form, section, and date — so its vintage is always visible. When newer filings arrive, a re-run produces a fresh, fully cited verdict against them; nothing is silently rewritten.

Three paid plans — Greenhorn ($29/month), Prospector ($69/month), and Mother Lode ($149/month) — with annual billing at ten months’ price, plus a free tier for browsing, portfolios, and the quantitative pre-screens. Paid plans include a 14-day first-time trial and cancel anytime. Independent of the plan, every scan shows its estimated processing cost before it runs and reports what it actually spent — cost transparency is part of the product. See https://www.rockturners.com/pricing/ for full details.

Yes, within limits. The free tier covers browsing the platform, keeping portfolios, and running the quantitative pre-screens — the Stage 1 filters that need no AI compute. The AI analyses — deep dives, cited verdicts, the Strategy Lab’s Stage 2 runs, and alert screens — come with a paid plan, each of which includes its full price as monthly compute. Every scan shows its estimated cost before it runs; nothing is ever charged automatically.