SEC Fundamentals Screener — XBRL Ratios Across All Filers

Rank every SEC filer by a fundamental ratio, computed from the official XBRL data the companies themselves filed. Not a list of filings — a ranking.

One request to SEC's XBRL frames API returns a single accounting figure for every filer in a period (total assets for Q4 2025 came back with 6,163 companies). A ratio therefore needs only a handful of requests to cover the whole market.

Ratios

Ratio	Formula
Net margin	net income ÷ revenue
Gross margin	gross profit ÷ revenue
Operating margin	operating income ÷ revenue
ROE	net income ÷ shareholders' equity
ROA	net income ÷ total assets
Asset turnover	revenue ÷ total assets
Debt to equity	total liabilities ÷ equity
Current ratio	current assets ÷ current liabilities
Cash to assets	cash ÷ total assets
Cash flow margin	operating cash flow ÷ revenue
Accrual ratio	net income ÷ operating cash flow — far from 1 means the earnings are not backed by cash

Or set your own numerator and denominator from: revenue, netIncome, operatingIncome, grossProfit, operatingCashFlow, assets, liabilities, equity, currentAssets, currentLiabilities, cash, longTermDebt, shares.

What you get per company

Field	Example
entityName / cik	ABBOTT LABORATORIES / 1800
ratio	0.1842
numerator / denominator	8,243,000,000 / 44,750,000,000
numeratorTag / denominatorTag	which US-GAAP tag the value actually came from
numeratorPeriodEnd / denominatorPeriodEnd	2025-12-31 / 2025-12-31
numeratorAccession / denominatorAccession	the filings the numbers came from

Two dataset views: Ranking and Provenance (tags and filing accession numbers, so every number traces back to a filing).

The two problems that make naive screens wrong

One accounting concept is spread across several US-GAAP tags. Measured for full-year 2025 revenue:

Tag	Companies
RevenueFromContractWithCustomerExcludingAssessedTax	2,705
Revenues	2,251
RevenueFromContractWithCustomerIncludingAssessedTax	658
Total	5,614

Using only the largest tag covers 48% of filers that reported revenue — and the output looks like a complete market screen either way. This Actor walks a fallback list per concept and records in every row which tag supplied the value.

A quarterly frame mixes fiscal calendars. For the same SEC period CY2025Q4, one company's net income covers 1 Sep – 30 Nov while another's revenue covers 1 Oct – 31 Dec. Divide them and the ratio spans different periods. This Actor compares the period end dates of same-type concepts, drops the mismatches, and counts them in SUMMARY.droppedPeriodMismatch. Annual periods pair about 5.8 times more companies than quarterly ones, which is why CY2025 is the default.

How it reports what it could not do

Check	What it catches
Concept unavailable	Every fallback tag returned 404 for that period. Reported as "no data for this concept and period", never as "no companies matched". Balance-sheet concepts need the I suffix on the period, and a mismatch is a hard 404.
Period mismatch	Numerator and denominator reported for different period end dates. Dropped, counted, and the run stops if most companies are affected.
Balance-sheet date far from the income period	For ROE, ROA and other income ÷ balance-sheet ratios, the balance-sheet date must be within maxInstantGapDays (default 10) of the income period's end. Otherwise a July fiscal year's income would be divided by January equity. Dropped and counted in droppedInstantMisaligned.
Low coverage	Too few filers can be paired. A ranking built on a small intersection is not the market, so the run fails loudly rather than looking complete.
Zero denominator	Dropped and counted, never returned as infinity.
Tiny denominator	minDenominatorAbs removes small absolute denominators. minDenominatorToAssets (default 0.002) removes denominators that are tiny next to the company's total assets: measured for CY2025 net margin, the unfiltered top two were a SPAC with $58 of revenue (5,212x) and a fund whose revenue tag was 0.009% of its assets. Mortgage REITs, the lowest legitimate revenue-to-assets values (0.006–0.008), stay in.
Misfiled scale	scaleCheck drops numerators larger than maxScaleRatio × total assets (a proxy statement filed net income 1,000 times too large, and SEC's frame picked it) and companies with total assets under minScaleAbs.
Missing contact	SEC requires programmatic requests to identify the requester, so the run is refused with the required format instead of sending an anonymous request. The input form's placeholder (Your Name your@email.com) and reserved addresses such as @example.com count as missing: they are refused before any request reaches SEC.

SUMMARY also carries the ratio distribution (min, p25, median, p75, max) and per-tag coverage, so you can see the shape of the market before reading the top of the list.

Notes on the data

• SEC's fair-access policy requires a contact name and email in the request header. You provide it; the Actor will not invent one.
• Values are taken as filed, with no restatement adjustments. Outliers are removed only by the scale checks above, and every removal is counted in SUMMARY.
• coverageOfLargerFrame tells you what share of filers this ranking represents. A low value means the ranking is a subset, not the market.
• SEC's annual balance-sheet frame holds one value per company near the calendar year end. Companies whose fiscal year ends elsewhere therefore drop out of ROE-type ratios rather than being divided by a balance from another point in their year.
• This is a screening tool, not investment advice.