Fundamental Data & Stock Research Platforms

Every product here will show you a revenue line for the same company, and they do not all mean the same thing by it. Three questions separate them and none is answered on a pricing page — where the numbers come from, whether restatements were folded back into prior years, and how many of those years the tier you are looking at actually unlocks. History depth is sold as a tier far more often than it is a property of the product: the same terminal can cost $24.95 or $119.95 a month depending only on how far back it will show you.

• Global coverage and deep history in a terminal — TIKR, on S&P Global Capital IQ data, from 5 years free to 30 on Ultimate.
• Five years of statements with no account at all — Stock Analysis.
• Ten years of US statements with no account and no ads — QuoteLinks, every figure from one provider, EODHD.
• The numbers in Excel or Google Sheets — Wisesheets at $60 a year.
• Segment and KPI data lifted out of the filing — Daloopa.
• US filings, free and keyless — SEC EDGAR, the source everything else here reads from.
• Filings, transcripts and investor presentations — Quartr, free on mobile.
• What changed in a company's risk factors since last year's 10-K — RiskDiff, 434 US filers, $199 a month for the company diffs.
• A valuation model whose inputs you can change — Stock Unlock for one DCF over seven metrics, Finbox for a gallery of models that export with their formulas.
• Screening across a long history of ratios — GuruFocus.

What actually decides the choice here

Every product in this listing will show you a revenue line for the same company, and they do not all mean the same thing by it. Seven questions separate them, and none is answered on a pricing page.

Where the numbers come from. There are three routes and they fail differently. A vendor can licence a standardised feed — TIKR names S&P Global Capital IQ, Finbox and Stock Analysis name S&P Global Market Intelligence, which is why their income statements agree with each other and not with the filing. It can tag filings itself, in which case the figures match the 10-K and the line items drift from company to company. Or it can ride the free floor: the SEC's XBRL company-facts endpoint, which needs no key and costs nothing, and which is the reason a dozen cheap products exist at all. Wisesheets API is the plainest case of the third route — US fundamentals taken from EDGAR XBRL rather than bought from an aggregator, with every value carrying the XBRL tag, accession number and filing date it came from. Equibles adds a fourth: an AGPL core that ingests the public records into a database on your own hardware, beside a hosted portal, REST API and MCP server that add managed datasets. A standardiser is slow and opinionated, a self-tagger is idiosyncratic and occasionally wrong, and the free endpoint silently drops anything a company tagged with its own extension elements. Ask which one you are buying.

How far back, and whether that is the product or the tier. History depth is sold by the tier more often than it is owned by the product. TIKR is the clearest case: the same terminal gives 5 years free, 10 years on Plus at $24.95 a month, and 30 years on Ultimate at $119.95; Pro, at $54.95, is sold as 20 years, of which the main S&P dataset tables only the first 10. Nothing about the underlying data changed between those screens. Stock Analysis shows why: its free five years are an S&P licence term, and the "10-40 years" on its Pro page is a range across providers, reached by switching the data source and accepting another vendor's line items. Read the advertised figure twice — Wisesheets advertises up to 30 years while its own Statement Dump documentation says 19. Decide first how many full cycles you need to see — for most of what people value on these platforms, ten years is less than one.

As-reported or restated. A vendor that folds restatements, discontinued operations and reclassified segments back into prior years gives you a clean, comparable series that nobody ever saw. A vendor that keeps the original figures gives you a messy series that is exactly what the market was pricing. Both are legitimate; only one is honest to backtest against, and the distinction is invisible in the interface. In XBRL the default convention makes it worse, because a restated amount is the plain value and the superseded one is the specially tagged exception. Where a product lets you choose — Stock Analysis toggles standardised against as-reported, and Wisesheets API takes asReported=true to return the raw XBRL facts — choose deliberately.

Segment and KPI data. The numbers that actually move these businesses — subscribers, bookings, same-store sales, remaining performance obligations, margin by segment — do not exist in any standardised chart of accounts, because they are different in every company. They are reported under company-specific extension tags, which the free SEC endpoint excludes by design, and a standardiser has nowhere to put them. Everything else in this category is available somewhere for $60 a year; this is the part that is genuinely hard, and it is the entire argument of the products built around it. Daloopa extracts them from filings, decks and supplementals for 6,000+ tickers and quotes every plan through sales; Fiscal.ai carries segments and KPIs for the largest 2,300 companies at a published price, with two years of KPI data free and ten on Pro at $49 a month.

Non-US coverage. This collapses faster than anything else and is where these products most differ in practice. GuruFocus is unusually candid about why — its US membership is $549 a year and Europe, Asia and Canada each cost extra, because each one had to be bought. Several products stop at North America outright: YCharts carries US and Canadian equities and nothing else at any tier, and FAST Graphs sells the rest as a $10-a-month add-on without publishing which exchanges it covers. Wisesheets API is SEC filers only. Even a product that carries a market may carry less of it: FinancialFilings records that IFRS filers get 65 of its 126 standardised line items against 91 for US GAAP. Most products are less explicit and leave you to discover the gap after you have built a watchlist. Read the exchange list, and treat "global" with the suspicion it has earned.

Whether the data leaves the product. A number you cannot export is a number you cannot check. Some products here are deliberately screenshot-only — Stock Unlock documents no export at all, and TIKR has no API or add-in at any price and exports to Excel only from Pro. Others exist for the opposite reason — Wisesheets sells financials across more than 50 exchanges into Excel and Google Sheets for $60 a year, or $120 for Elite, which adds options data and an unquantified "faster data speeds", precisely because it sells the data rather than an interface to look at it in. Its sibling, Wisesheets API, sells the same SEC fundamentals over REST and a hosted MCP server, free for five years of history and then $19 or $39 a month billed annually. Check what the export actually contains: Stock Analysis's downloads come from a fallback provider because S&P bans exports, so the file can disagree with the table on screen. If your work ends in a model of your own, put export near the top of the list, and find out whether that means an API, a CSV, a spreadsheet function or a PDF.

How much of the screen is the vendor's own model. Fair values, quality grades, composite scores, normalised owner earnings and predicted returns are outputs, not observations. They are often the most prominent figure on the page and the least attributable. GuruFocus's GF Value has no published formula, and YCharts prints an Attractive, Neutral or Avoid Y-Rating beside its data. The usable ones tell you the formula and let you change the inputs; the rest are a conclusion wearing the typography of a fact.

The words this market uses loosely

"Fundamental data" covers three different things — the figure as reported, the figure after a vendor mapped it onto a common template, and a figure the vendor computed. Products move between the three without marking the boundary.

"Fair value" and "intrinsic value" are never properties of a company. They are the output of a discount rate, a growth path and a terminal assumption that somebody chose. The only question worth asking is whose, and whether you can edit them.

"As-reported" should mean the original filing, untouched by later restatement. Some vendors use it to mean only that line items were not remapped, while still carrying restated values.

"KPI" and "segment data" are used both for figures lifted out of the filing and for figures a vendor derived from them. Derived numbers are fine; being unable to tell which is which is not. This is why the serious KPI products hyperlink every cell back to the page it came from.

"Coverage" is three unrelated numbers wearing one word — how many tickers, how far back, and how many line items per company. A vendor quoting 100,000 companies may carry five years and forty line items for most of them.

Three jobs sharing one audience

Fundamentals is the data job: assembling a history you can rely on. The buyer wants source, depth, line-item count, restatement policy and export, and does not care what the interface looks like. Judge these against each other on the seven questions above, and on nothing else. The interface still decides who the product is for. Fiscal.ai is a terminal with 20+ years of annual financials and an AI Copilot that links each number back to the filing page it was read from, and its API and MCP server are a separate purchase with a free key capped at 100 companies. Godel Terminal puts standardised financials, estimates and SEC filings behind Bloomberg-style commands with real-time US quotes, at $118 a month and with no API for an individual. QuoteLinks is the free end of the same job: ten fiscal years of statements, estimates and 13F holders for US-listed stocks, no account and no ads, with EODHD's figures passed through unchecked — on the day we read it, its screener listed a micro-cap at a $299 billion market cap. YCharts is built for advisers turning research into client documents, North American equities only, and quotes every tier through sales. Equibles is the one whose core you can run yourself.

Valuation is the modelling job: turning that history into a number you would act on. The product is a set of assumptions and the freedom to change them, so judge it on which models it offers, which inputs are editable, whether it shows you the sensitivity rather than a single figure, and whether it validates against enough history to be worth running. A valuation tool that will not let you move the discount rate is a rating service with extra steps. That is where the line sits between an instrument and a rating, and it runs through this listing rather than around it: a fair value you can re-derive and whose inputs you can change is an instrument, one you cannot is a rating. Finbox prints a headline fair value on every stock page, but every DCF, multiples and dividend discount model behind it opens with the assumptions editable and exports to Excel or Google Sheets with the formulas intact. FAST Graphs labels its orange line fair value, and its own documentation says that line is a P/E of 15 for most companies, a multiple you can replace with your own. Stock Unlock hands over the growth rates, discount rate, terminal multiple and projection window of its DCF. GuruFocus's GF Value fails the test, and the product is listed for its 30 years of financials and its screener, not for that number.

Filings is the documents job: 10-Ks and 20-Fs, transcripts, presentations, 8-Ks, and getting to them quickly. Judge on latency after the call ends, how far the transcript archive goes back, whether non-US investor-relations material is covered at all, and whether the text is searchable across companies rather than one document at a time. Transcript history is another thing routinely sold by the tier — TIKR gives 90 days free, ten years on Pro and the full history on Ultimate. SEC EDGAR is the floor: free, keyless, ten requests a second, full-text search back to 2001, and no transcripts at all. CapEdge puts boolean search over filings and earnings-call transcripts, 10-K diffs and email alerts on top of it for nothing, with no API and terms that forbid scraping. RiskDiff narrows the diff to Item 1A: each risk factor matched to last year's, scored 1 to 10 and tagged by theme for 434 US filers, from $29 a month for the cross-company theme index and $199 for the company diffs, with an MCP endpoint on both. BamSEC does the document work more carefully — labelled filings, extracted key exhibits, any table to Excel in one click, transcripts beside the filings — for $69 a month billed annually, and also has no API. sec-api.io is the same corpus as JSON, with a websocket stream of new filings, from $55 a month, though full-text search is not on that tier. FinancialFilings is the route outside the SEC: filings from 60 markets converted to Markdown, a free web index, 500 API credits a month free and $0.01 a credit after. For the call itself, Quartr's free app carries live audio, transcripts and decks for 16,000-plus companies, with transcripts only for events held in English; Quartr API is the same material on a quote-only annual contract with written delivery times; and EarningsCall API is the self-serve route to transcripts as data, from $60 a month for personal use and $80 for anyone using it for an employer, with most companies starting in 2020.

The three share an audience and almost nothing else. A KPI extraction service, a DCF workbench and a transcript archive answer different questions, and a ranking that puts them in one column is measuring nothing. Decide which job you are here for, then compare only inside it.

One filter worth applying yourself

A search for the best fundamental analysis site returns two markets mixed together, and they are judged on completely different things. Switch the vendor's opinion off and see what remains. If a filing archive, a financial history and a model you control are left, it is an instrument, and coverage, depth and export are the right questions. If the star rating or the analyst fair value was the reason to subscribe, it is a recommendation service, and the only question that matters is its track record. This catalogue lists the first kind, which is why several well-known names are absent from the listing below. Some listed products still carry a verdict — GuruFocus's GF Value, YCharts' Y-Rating, Stock Unlock's 1–5 scores — and are listed for what is left when it is switched off.

Where this category connects

If you want these numbers in your own code rather than in an interface, several of the providers here sell an API and are listed as such in market data APIs, where ROIC.ai's fundamentals API has its primary place. Several products in the listing below are primary elsewhere and appear here as secondary members: Koyfin, Stock Rover, Stockopedia and Screener.in are bought first as screeners and live in charting and screeners, and AlphaSense, BamSEC's parent, is an enterprise research copilot in AI agents, as is DeepVal, whose AI-written company reports are the whole product. Form 4 and 13F data as a job of its own is in insider and 13F trackers. A valuation history that a strategy will be tested against has a further requirement nothing here guarantees — point-in-time figures free of survivorship bias — which is the first thing to settle in backtesting frameworks.