# StockFit API

> StockFit is a developer API for accurate US stock fundamentals and ETF data, parsed directly from SEC EDGAR. Endpoints cover SEC filings (10-K, 10-Q, 8-K, S-1, 13F, DEF 14A, Forms 3/4/5, NPORT-P), structured financials (income statement, balance sheet, cash flow), earnings history, executive compensation, institutional and insider ownership, ETF holdings and flows, and AI-classified business model analytics. All endpoints are also available as MCP tools so the same API key works with Claude, Cursor, and VS Code.

The StockFit API parses raw XBRL, XML, and SGML directly from SEC EDGAR. There are no third-party aggregators, no normalized or "adjusted" metrics that hide source data — every value is traceable back to the underlying filing. The same REST API is exposed as a Model Context Protocol (MCP) server so AI assistants can call endpoints as tools using the same API key.

Company: Second Dot LLC. Brand: StockFit. Contact: info@stockfit.io.

## Product pages

- [Landing page](https://developer.stockfit.io): Overview of all API capabilities, live playground, and FAQ covering each endpoint family.
- [For Quants](https://developer.stockfit.io/quant): Landing page aimed at quantitative researchers and systematic traders. Point-in-time SEC fundamentals, filing-timestamped values, normalized statements, no look-ahead or survivorship bias.
- [For App Developers](https://developer.stockfit.io/app-developers): Landing page for developers building investing apps, websites, and dashboards. Normalized company fundamentals on one schema, batched OHLCV price quotes and historical bars for charting, dividend history with payout ratio and coverage, forward and reverse stock splits, executive profiles and compensation, company lookup by ticker/CIK/CUSIP/FIGI, and ETF holdings. Includes live coverage counts.
- [For Hedge Funds](https://developer.stockfit.io/hedge-funds): Landing page for institutional investors and fund managers. Point-in-time company fundamentals with per-value restatement trails, 13F institutional ownership aggregated by manager with prior-quarter deltas, Schedule 13D/13G beneficial owners, Form 3/4/5 insider transactions, executive compensation and governance, and the full SEC filings corpus. Survivorship-bias-free: delisted companies keep their complete history, resolvable by CIK, CUSIP, or FIGI.
- [Economic Model / Qualitative Business Insights](https://developer.stockfit.io/economic-model): Landing page for the qualitative company research API. Per-company flywheels (ordered self-reinforcing loops labeled by whether they compound margin, defensibility, or growth), failure modes with transmission mechanisms and watch metrics, operating levers with directionality, structural advantages, strategic initiatives, unit economics, and capital allocation. Extracted from SEC 10-K, 10-Q, and DEF 14A filings only, with a form type, item section, verbatim quote, and filing URL behind every claim.
- [ETF & Fund Data](https://developer.stockfit.io/etf): Landing page for the ETF and mutual fund API. Daily holdings published by issuers every trading day plus full quarterly N-PORT portfolios, portfolio overlap between any two funds measured from both sides, reverse lookup of every fund holding a given stock, fund flows with monthly granularity, expense ratios and fee schedules, brokerage commission and in-kind creation analysis, tracking difference against the index before and after fees, fair-value hierarchy, and service-provider rosters. Sourced from SEC N-PORT, N-CEN and N-1A filings.
- [Pricing](https://developer.stockfit.io/pricing): Free, Starter ($15/mo), Stock ($39/mo), ETF ($39/mo), and Professional ($69/mo) plans. Annual billing available. Free tier requires no credit card.
- [MCP + AI Agents](https://developer.stockfit.io/mcp): Two-step setup to connect StockFit to Claude Desktop, Claude Code, Cursor, or VS Code via the Model Context Protocol. The same API key works as a REST key and an AI key.
- [Playground](https://developer.stockfit.io/playground): Interactive endpoint explorer.

## API documentation

- [API documentation (embedded)](https://developer.stockfit.io/docs): Same Scalar OpenAPI reference embedded inside the developer portal — keeps users on developer.stockfit.io while browsing the API surface. The underlying source is the api.stockfit.io reference below.
- [Full Swagger documentation](https://api.stockfit.io/docs): Complete OpenAPI reference for every endpoint, grouped by tag (financials, earnings, sec-filings, etfmf, ownership, executives, company, insider-transactions, symbol-lookup). Includes a tier-filter dropdown so users can narrow the view to endpoints available on a specific subscription plan.

### Tier-scoped API references (for coding agents)

Each tier page lists only the endpoints callable on that plan. When prompting a coding agent (Cursor, Copilot, Claude Code, etc.) to write code against the StockFit API, reference the URL matching the user's subscription so generated code stays within the plan — calls outside the listed set return HTTP 403. MCP clients don't need these URLs because the MCP server already filters available tools by the user's tier server-side.

- [StockFit API — Free tier reference](https://api.stockfit.io/docs/free): Tier-scoped API reference listing only the endpoints accessible on the Free plan (no credit card). Reference this URL when prompting a coding agent to write API code for a free-tier account.
- [StockFit API — Starter tier reference ($15/mo)](https://api.stockfit.io/docs/starter): Tier-scoped API reference for the Starter plan. Adds earnings snapshots, date predictions, EPS history, financial growth rates, and the filings stream on top of the free tier.
- [StockFit API — Stock tier reference ($39/mo)](https://api.stockfit.io/docs/stock): Tier-scoped API reference for the Stock plan. The full stock fundamentals workflow: earnings calendar, multi-year trends, 8-K Item 2.02 body extraction, insider transactions, ownership, executives.
- [StockFit API — ETF tier reference ($39/mo)](https://api.stockfit.io/docs/etf): Tier-scoped API reference for the ETF plan. Fund holdings, flows, overlap, fee analysis, service providers, exposure models, and the N-PORT / N-CEN power tools.
- [StockFit API — Professional tier reference ($69/mo)](https://api.stockfit.io/docs/pro): Tier-scoped API reference for the Professional plan. Every endpoint across stock fundamentals, ETF analytics, ownership, insider transactions, executives, and AI-classified company analytics.

### Endpoint families

- SEC Filings: `/v1/api/filings`, `/v1/api/filings/item`, `/v1/api/filings/item-list`, `/v1/api/filings/latest`, `/v1/api/filings/timeline`, `/v1/api/filings/calendar`, `/v1/api/filings/stats`, `/v1/api/filings/recent-offerings`, `/v1/api/filings/offering`. Covers 10-K, 10-Q, 8-K, S-1, DEF 14A, Forms 3/4/5, NPORT-P, N-CEN, and more.
- Financials: `/v1/api/financials/income-statement`, `/v1/api/financials/balance-sheet`, `/v1/api/financials/cash-flow-statement`, `/v1/api/financials/as-reported`, `/v1/api/financials/key-metrics`, `/v1/api/financials/growth`, `/v1/api/financials/scores`, plus chart variants (`/v1/api/financials/chart/...`). All values from raw XBRL.
- Earnings: `/v1/api/earnings/calendar`, `/v1/api/earnings/date`, `/v1/api/earnings/snapshot`, `/v1/api/earnings/eps-history`, `/v1/api/earnings/dividend-history`, `/v1/api/earnings/trends`, `/v1/api/earnings/chart/eps`, `/v1/api/earnings/chart/quality`.
- ETF / Mutual Fund: `/v1/api/fund/holdings`, `/v1/api/fund/holdings/daily`, `/v1/api/fund/holdings/daily/supported-funds`, `/v1/api/fund/composition`, `/v1/api/fund/overlap`, `/v1/api/fund/reverse-lookup`, `/v1/api/fund/flows`, `/v1/api/fund/fees`, `/v1/api/fund/fee-analysis`, `/v1/api/fund/exposure-model`, `/v1/api/fund/performance`, `/v1/api/fund/profile`, `/v1/api/fund/structure`. Holdings come in two cadences: full quarterly portfolios from SEC NPORT-P filings, and daily holdings published on each fund's own website (updated every trading day).
- Ownership: `/v1/api/ownership/institutional-holders` (13F), `/v1/api/ownership/portfolio`, `/v1/api/ownership/beneficial-owners` (Schedule 13D/13G), plus history endpoints. Insider transactions (Form 3/4/5) at `/v1/api/insider-transactions`, `/v1/api/insider-transactions/by-insider`, `/v1/api/insider-transactions/roster`, `/v1/api/insider-transactions/summary`.
- Congress Trades: STOCK Act periodic transaction reports (PTRs) of U.S. House members and Senators, served as disclosed (amount ranges, owner code, amendments, source filing URL) with best-effort links to the traded company. Per chamber: `/v1/api/congress/house/trades` and `/v1/api/congress/senate/trades` (by company), `/v1/api/congress/house/trades/by-member` and `/v1/api/congress/senate/trades/by-member` (by Bioguide ID), `/v1/api/congress/house/trades/recent` and `/v1/api/congress/senate/trades/recent` (market-wide feed of new filings), `/v1/api/congress/house/members` and `/v1/api/congress/senate/members` (roster of filers).
- Executives: `/v1/api/executives/compensation` (pay-vs-performance from DEF 14A), `/v1/api/executives/officers`, `/v1/api/executives/governance`, `/v1/api/executives/performance-measures`.
- Company: `/v1/api/company/details`, `/v1/api/company/peers`, `/v1/api/company/economic-model`, `/v1/api/company/research-summary`.
- Symbol Lookup: `/v1/api/lookup/symbol`, `/v1/api/lookup/cik`, `/v1/api/lookup/cusip`, `/v1/api/lookup/figi`, `/v1/api/lookup/batch`, `/v1/api/lookup/former-names`, `/v1/api/lookup/news`, `/v1/api/lookup/search`.

## Developer blog

Technical guides and engineering notes for developers building on the StockFit API.

- [All posts](https://developer.stockfit.io/blog): Browse the full developer blog.
- [ETF Turnover Rate, From Daily Data](https://developer.stockfit.io/blog/etf-turnover-rate-api): Explains the standard portfolio turnover ratio formula (lesser of purchases or sales divided by average net assets, Form N-1A Item 3) and shows /api/fund/changes only diffs quarterly N-PORT filings, confirmed for SPY, QQQ, JEPI, TQQQ, SOXL, and ARKK, a two-to-three-month gap between previousDate and latestDate every time, not a daily turnover feed. Walks through computing an annualized turnover estimate directly from two /api/fund/holdings/daily snapshots (2026-08-10 to early September 2026) for the same 6 funds, with curl and JSON examples, matching positions by cusip, ticker, then name. Surfaces two real data-quality findings: SPY's daily file has no valueUsd field at all (weight-only pctVal-based turnover computed instead) and SOXL's swap counterparty identifiers get reformatted between snapshots, inflating naive turnover to 873.1% before reconciling 8 of 9 pairs down to 47.0%. Cross-checks the computed estimates against three funds' own primary-sourced official turnover ratios: SPY 3% (SEC EDGAR 485BPOS, fiscal year 2024), ARKK 43% (ARK's own summary prospectus), and JEPI 172.17% (JPMorgan's own fact sheet, itself over 13 months stale relative to its own cover date), showing the computed-versus-official gap narrows from 15x to 2x to 1.4x as real trading activity rises. Two Highcharts charts (turnover by fund type, computed versus official). Endpoints: /api/fund/changes and /api/fund/holdings/daily (ETF and Professional plans), /api/fund/holdings (Starter plan and above), /api/fund/holdings/daily/supported-funds (free tier). 7-question FAQ covers what counts as a good turnover ratio, how to compute it from holdings data, why /api/fund/changes is not daily, why computed estimates run high, official reporting cadence, N-PORT versus daily holdings, and required subscription tier.
- [How Often Do ETF Holdings Actually Change?](https://developer.stockfit.io/blog/how-often-etf-holdings-update): Data-story built from /api/fund/holdings/daily, diffing two real daily holdings snapshots (2026-08-10 to early September 2026, roughly 25 days apart) for 10 funds: four index trackers (SPY, IVV, QQQ, DIA), three active funds (ARKK, JEPI, JEPQ), and three leveraged/inverse funds (TQQQ, SOXL, SQQQ), matched position by CUSIP with ticker and name fallback. Splits the change into two signals: membership change (a ticker added or dropped) versus balance resizing (the same ticker, a different share or notional count). Findings: SPY and IVV both track the S&P 500 yet moved 1.6% and 99.6% of positions respectively over the same window, because IVV's swing was a fund-flow-driven near-uniform 3.164% share reduction across 504 of 506 positions, not stock-picking; the one real reconstitution event (Reddit/RDDT added, AvalonBay/AVB removed after merging with Equity Residential into Vivmark Residential/VMRK on 2026-08-17/18) reads as 1 added/1 removed in SPY's CUSIP-keyed file but 2/2 in IVV's file, which carries no CUSIP field at all; ARKK added Rocket Lab (RKLB) and exited Deere (DE), Roblox (RBLX), and Snowflake (SNOW); leveraged funds TQQQ and SOXL resize 95%+ of positions daily (mostly index total-return swaps against Citibank, BNP Paribas, and Bank of America) to hold their multiple while genuine new counterparties stay under 3%; JEPQ's apparent 18% ticker churn is entirely scheduled options-note rolls with zero change to its underlying stock roster. Separately quantifies Form N-PORT's public disclosure lag using SPY's (SPDR S&P 500 ETF Trust, CIK 884394) last five NPORT-P filings cross-checked against SEC EDGAR directly: 57 to 59 days after each quarter close, versus a daily file refreshed within days. Covers how to pin the reportDate parameter to a historical date and read the response's own reportDate field to confirm what was actually served, with real /api/fund/holdings/daily curl and JSON examples. Two Highcharts charts (cadence by fund category, membership-versus-resize composition) plus an N-PORT filing-lag table. Endpoints: /api/fund/holdings/daily (ETF and Professional plans), /api/fund/holdings/daily/supported-funds (free tier), /api/fund/holdings (Starter plan and above). 8-question FAQ covers update frequency, whether a new file means a trade happened, index versus active turnover, why leveraged ETFs resize without adding holdings, N-PORT staleness, pinning a historical reportDate, daily versus N-PORT holdings, and required subscription tier.
- [Financial Health Score API: Point-in-Time Data](https://developer.stockfit.io/blog/financial-health-score-api): Developer walkthrough and data-story on /api/financials/scores and /api/financials/chart/scores, which compute the Piotroski F-Score (0-9, nine year-over-year tests across profitability, leverage/liquidity, and operating efficiency, with the individual pass/fail detail on piotroskiDetails) and the Altman Z-Score (the classic five-ratio bankruptcy-prediction model: working capital, retained earnings, EBIT, equity, and revenue, each scaled by assets or liabilities, sorted into safe/grey/distress zones at the 2.99 and 1.81 thresholds from Altman's original 1968 Journal of Finance paper) directly from each filer's own standardized, point-in-time financial statements, never recomputed on today's restated numbers. Manual verification walkthrough reproduces Boeing's (BA, CIK 12927) FY2025 Altman Z-Score by hand from the raw balance-sheet and income-statement facts (accession 0001628280-26-004357, filed 2026-01-30): computed 0.9467, matching the API's 0.95 to two decimals. Worked example charts Boeing's Piotroski and Altman scores across fiscal 2016-2025, both collapsing together in fiscal 2019 (the year of the FAA's March 13, 2019 737 MAX grounding), bottoming further in the fiscal-2020 COVID year, and the Altman Z-Score remaining in the model's distress zone every year since, including a fiscal-2024 Piotroski dip tied to the January 2024 Alaska Airlines door-plug incident and the fall 2024 machinists' strike, despite Boeing never missing a debt payment. Cross-sectional comparison across 8 companies' most recent fiscal year finds Apple (AAPL) and Microsoft (MSFT) both landing in the model's grey zone rather than safe (a 1968 manufacturing-era calibration artifact, not a credit-risk signal), and Ford (F), Carnival (CCL), Delta Air Lines (DAL), and Boeing all reading distress-zone as a structural pattern of capital-intensive leverage rather than four companies near default. Compares against Financial Modeling Prep's own Financial Health Scores API with appropriately hedged, publicly-documented-only claims, and links to the fuller StockFit-vs-FMP pricing and endpoint comparison. Two Highcharts charts built from that captured data. Endpoints: /api/financials/scores and /api/financials/chart/scores (Starter plan and above), /api/financials/balance-sheet and /api/financials/income-statement (free tier, used for manual verification). 8-question FAQ covers what a financial health score is, how to call the API, what counts as a good Altman Z-Score, why Apple and Microsoft don't score safe, backtesting without lookahead bias, Piotroski vs Altman, whether a low Z-Score proves bankruptcy risk, and required subscription tier. Not investment advice.
- [Pension Funded Status API](https://developer.stockfit.io/blog/pension-funded-status-api): Product walkthrough and data-story on /api/footnotes/retirement-plans, which returns the full ASC 715 defined-benefit pension and OPEB (other postretirement benefits, primarily retiree healthcare) footnote per filer: funded status (plan assets minus benefit obligation), benefit obligation, plan-asset book cross-tabbed by category and fair-value level (Level 1/2/3/NAV), cost components (service cost, interest cost, expected return), employer contributions, and discount-rate and expected-return assumptions, each cited to dateFiled. Explains why pension and OPEB funded positions have no dedicated balance-sheet line and instead fold into broader asset or liability categories, shown live on Lockheed Martin's (LMT) FY2025 10-K balance sheet, where the pension deficit lands in otherNonCurrentLiabilities. Cross-company funded-ratio snapshot across nine industrials, airlines, and defense contractors (FY2025, period 2025-12-31): Lockheed Martin (LMT) the most underfunded pension at 85.5% ($3.9 billion deficit, narrowing three years straight from $6.2 billion in FY2023), Honeywell (HON) the most overfunded at 134.7% ($5.3 billion surplus), also covering Boeing (BA), UPS, American Airlines (AAL), Ford (F), GE Aerospace (GE, with a note on its April 2024 GE Vernova/GE HealthCare corporate split), Northrop Grumman (NOC), and Delta Air Lines (DAL). Second data-story finding: retiree healthcare (OPEB) is funded far worse than pension almost everywhere, five of seven companies under 10% funded, because ERISA's minimum-funding mandate does not reach OPEB (sourced to the Department of Labor's own ERISA guidance), except two defense contractors, Lockheed Martin (161.0%) and Northrop Grumman (113.3%), whose OPEB is funded better than their own pension. Two Highcharts charts built from that captured data. Endpoints: /api/footnotes/retirement-plans (Professional plan), /api/financials/balance-sheet and /api/footnotes/debt-structure (free tier and Professional plan respectively). 8-question FAQ covers what funded status means, how to call the API, why the balance sheet hides it, pension vs OPEB, why OPEB is rarely pre-funded, discount-rate assumption changes, which companies are underfunded in 2025, and required subscription tier. Not investment advice.
- [Operating vs Finance Lease API: ASC 842 Data](https://developer.stockfit.io/blog/operating-finance-lease-api): Developer walkthrough and data-story on /api/footnotes/leases, which returns the full ASC 842 lease footnote per filer: right-of-use asset, lease liability (current and noncurrent, split by operating and finance classification), the undiscounted payment maturity ladder (year 1 through year 5 plus thereafter), weighted-average discount rate, and the leaseNotYetCommenced disclosure for signed-but-unstarted leases. Explains the five ASC 842 classification tests that decide operating versus finance (ownership transfer, bargain purchase option, major-part-of-life, substantially-all-fair-value, specialized asset). Data story: leases not yet commenced (off-balance-sheet AI datacenter capacity commitments) across five hyperscalers, each cross-checked to the dollar against SEC's raw XBRL companyconcept API (FinanceLeaseLiability, OperatingLeaseLiabilityNoncurrent). Oracle (ORCL, CIK 1341439) grew from $43.4 billion to $260.0 billion in one fiscal year (499%), nearly double its $130.1 billion of funded debt. Microsoft (MSFT, CIK 789019) grew from $92.7 billion to $329.1 billion (255%), with an 8-quarter deep dive showing a real decline-then-inflection shape, not a straight line. Amazon (AMZN), Meta (META), and Alphabet (GOOGL) included; combined five-company total grew from $239.1 billion to $847.7 billion. Worked classification comparison: Walmart's (WMT, CIK 104169) lease liability is 30.3% finance versus Delta Air Lines' (DAL, CIK 27904) 11.6% finance, with full payment-ladder detail for both. Three Highcharts charts built from that captured data. Endpoints: /api/footnotes/leases (Professional plan), /api/financials/balance-sheet and /api/filings/search-by-accession-number (both free tier). 8-question FAQ covers operating vs finance classification, calling the API, leases not yet commenced, the AI datacenter lease signal, spotting a company's classification mix, the payment maturity schedule, lease liability versus funded debt, and required subscription tier. Not investment advice.
- [Remaining Performance Obligation (RPO) API](https://developer.stockfit.io/blog/remaining-performance-obligation-api): Developer walkthrough and data-story on /api/footnotes/backlog, which returns a filer's remaining performance obligation (RPO), the ASC 606-10-50-13 disclosure of contracted revenue not yet recognized, straight from the 10-K/10-Q XBRL tag us-gaap:RevenueRemainingPerformanceObligation, cited to the SEC accession number. Explains how RPO, backlog, deferred revenue, and current RPO (cRPO) differ, and shows the endpoint's expectedShareInWindow field via Adobe's FY2025 filing (65% of RPO expected to convert within the disclosed window). Worked comparison across three companies with clean six-year (FY2021-FY2026) histories, each cross-checked to the dollar against SEC's raw XBRL companyconcept API: Microsoft (MSFT, CIK 789019) RPO grew from $146 billion to $684 billion, an 82% jump in fiscal 2026 alone against 17.8% revenue growth (a 36.2% five-year RPO CAGR versus 14.6% revenue CAGR, RPO now 206% of annual revenue), explicitly distinguished from the $678 billion "commercial RPO" non-GAAP figure Microsoft's own earnings materials headline. CrowdStrike (CRWD, CIK 1535527) RPO grew 6.4x in five years ($1.4 billion to $9.0 billion), slightly outpacing its own hypergrowth revenue. Salesforce (CRM, CIK 1108524) keeps RPO and revenue growing in near lockstep around 14-15% CAGR, the steady-state control case. Two Highcharts charts built from that captured data: an indexed RPO growth trajectory (FY2021=100) across all three companies, and a grouped revenue-CAGR-vs-RPO-CAGR comparison. Endpoints: /api/footnotes/backlog (Professional plan), /api/financials/income-statement and /api/filings/search-by-accession-number (both free tier). 8-question FAQ covers what RPO discloses, how to call the API, RPO vs deferred revenue, why Microsoft's RPO jumped 82%, disclosure coverage (roughly 15% of active filers), spotting bookings-momentum acceleration, current RPO (cRPO), and required subscription tier. Not investment advice.
- [REIT Property Schedule API: Schedule III Data](https://developer.stockfit.io/blog/reit-property-schedule-api): Product walkthrough of /api/footnotes/reit-properties, which turns SEC Schedule III (Real Estate and Accumulated Depreciation, required under Regulation S-X Rule 12-28) into structured per-property JSON: initial cost split into land and buildings, current carrying amount, gross carrying value, accumulated depreciation, and capitalized improvements, sourced from the 10-K real estate exhibit via Arelle XBRL extraction, not the single netted real-estate line on a balance sheet. Worked end to end on EastGroup Properties (EGP, CIK 49600) FY2025 10-K, period 2025-12-31, filed 2026-02-11: all 469 property rows sum to $6.7 billion gross, cross-checked to the dollar against SEC's own raw XBRL frames API (data.sec.gov companyconcept, accession 0000049600-26-000010) and reconciled against EGP's own balance sheet (net real estate is 94.2% of total assets, consistent with a pure-play industrial REIT). Two charts built from that captured data: gross value by state (Texas 36.5%, Florida 20.7%, California 14.4%, three states holding 71.6% of the portfolio) and accumulated depreciation as a percent of gross value by state, a portfolio-vintage signal ranging from 5.2% in Tennessee to 67.4% in Louisiana. Covers how filers tag this schedule differently, real quoted examples from Ventas (VTR, operator-portfolio-level rows like Atria Senior Living) and Iron Mountain (IRM, individual street-address-level rows spanning six world regions), and a data-honesty section on why some REITs return an empty properties array even when they do disclose a real Schedule III, with a recommended dateFiled and balance-sheet reconciliation check before trusting any ticker's response. Endpoints: /api/footnotes/reit-properties, /api/financials/balance-sheet, /api/filings/search-by-accession-number. FAQ covers what Schedule III discloses, how to call the API, initial cost vs carrying amount, estimating portfolio age from accumulated depreciation, why coverage varies by filer, confirming Schedule III excludes occupancy and rental income, data freshness, and required subscription tier (Professional plan). Not investment advice.
- [How to Detect Institutional Buying and Selling](https://developer.stockfit.io/blog/detect-institutional-position-changes): Product walkthrough of the StockFit institutional-ownership (SEC Form 13F-HR) API for detecting accumulation and distribution across quarters. Covers /api/ownership/institutional-holders (holder list by ticker with prevShares and prevPctOfShares for one-call quarter-over-quarter change tracking), /api/ownership/institutional-holders/history (aggregate holder count, shares, and value per quarter), /api/ownership/portfolio (a manager's full holdings by CIK, with each filed CUSIP resolved to a ticker symbol automatically), and /api/ownership/summary. Live AAPL data (2025-03-31 through 2025-12-31) shows BlackRock accumulating its stake every quarter (7.80% to 7.92% of shares outstanding) while Vanguard rises then drops to a real distribution quarter before partially recovering (9.11% to 9.19% to 8.60% to 8.76%), plus a Berkshire Hathaway (CIK 1067983) portfolio example resolving CUSIP 191216100 to Coca-Cola (KO). Explains the three-way voting-authority breakdown (sole, shared, none) per SEC Form 13F instructions with a real Vanguard vs BlackRock vs State Street comparison chart, why Form 13F excludes short positions and cannot show a fund's true net exposure, how to detect a full position exit via a set-difference of manager CIKs across quarters (not a low-share row), and the difference between 13F institutional ownership, Schedule 13D/13G beneficial ownership, and Form 4 insider transactions. Verified competitive comparison on the accession-number audit trail: Financial Modeling Prep and Intrinio return CIK only with no accession number, sec-api.io carries an accession number only at the parent filing level, and Massive matches StockFit's row-level citation. Endpoints: /api/ownership/institutional-holders, /api/ownership/institutional-holders/history, /api/ownership/portfolio, /api/ownership/summary, /api/company/details, /api/filings/search-by-accession-number. Stock or Professional plan required for ownership endpoints; company details on the free tier. 8-question FAQ. Not investment advice.
- [Why 13F Filing Data Is Always 45 Days Stale](https://developer.stockfit.io/blog/13f-filing-data-lag): Data-story analysis of the SEC Rule 13f-1(a)(1) 45-day filing lag on Form 13F-HR institutional ownership data, built from every initial (non-amended) 13F-HR filing StockFit has indexed rather than secondhand citations. Primary dataset: the 2025-09-30 quarter, 8,206 initial 13F-HR filings, charting cumulative percent filed by calendar day after quarter end (only 13.3% filed by day 17, 51.2% by day 38, a jump from 73.4% to 94.1% between day 44 and the day-45 deadline) and a weekly-bin breakdown showing 20.7% of the entire quarter filing on the exact deadline day, more than any other single week. Cross-quarter consistency check across seven consecutive quarters (2024-06-30 through 2025-12-31) shows 28.7% to 33.6% of each quarter's filings landing in deadline week every time. Named case study: Berkshire Hathaway Inc (CIK 1067983) filed its Form 13F-HR on the exact legal deadline for seven consecutive quarters, including the 2025-12-31 quarter where the normal February 14, 2026 deadline (a Saturday) rolled to February 17, 2026 under Rule 0-3(a) because the following Monday was a federal holiday, matching the SEC's own Form 13F FAQ example and independently cross-verified live against the production API. Explains the mechanics: what Form 13F-HR discloses (long equity positions only, no short interest, $100 million discretionary threshold), why the 45-day window exists (front-running and copycat-trading deterrence), and how to query /api/ownership/institutional-holders with an explicit reportDate to get complete quarterly coverage instead of a thin, still-filling default. Endpoints: /api/ownership/institutional-holders, /api/ownership/institutional-holders/history, /api/filings/search-by-accession-number, /api/company/details. FAQ covers why the lag exists, when filings become public vs when a quarter is complete, the weekend/holiday deadline shift rule, Berkshire's filing pattern, whether 13F covers short positions, getting complete quarterly coverage, initial vs amended 13F-HR/A filings, and required subscription tier (Stock or Professional plan). Not investment advice.
- [AI Economic Model API: Verified Business Thesis](https://developer.stockfit.io/blog/ai-economic-model-api): Engineering deep dive on /api/company/economic-model, the AI-generated business-model analysis (offerings, monetization, cost structure, operating levers, structural advantages/moats, flywheels, failure modes, unit economics, capital allocation) built from a company's own 10-K, 10-Q, and DEF 14A, where every claim carries a clickable SEC EDGAR URL, a filing section, and a verbatim 20-400 character quote machine-verified (literal substring match against the fetched, HTML-stripped filing text) before it is ever stored; citations that fail the check are dropped rather than kept, and items with no valid source left are dropped entirely. Cites FinanceBench (arXiv:2311.11944, Patronus AI/Contextual AI/Stanford), which found GPT-4-Turbo with retrieval incorrectly answered or refused 81% of SEC-filing questions with hallucinations across every model tested, to frame why verification matters. Worked end to end on Amazon (AMZN) FY2025 10-K: two independently AI-derived flywheels (a margin loop and a growth loop, both distinct slices of the famous Bezos napkin sketch, sourced separately rather than copied from it), four failure modes with mechanisms and watch metrics, and a seven-segment revenue reconciliation (Online Stores, Third-Party Seller Services, AWS, Advertising, Subscription Services, Physical Stores, Other) cross-checked dollar for dollar against /api/footnotes/revenue-segmentation, plus a live strategic initiative sourced to Amazon's Q1 FY2026 10-Q (the $15 billion OpenAI Series C investment). Includes a step-by-step guide to Ctrl+F verifying any citation yourself directly on SEC EDGAR, a confidence-level comparison (Amazon high vs Apple medium, captured the same day), and the 202-generating/200-ready polling pattern. Endpoints: /api/company/economic-model, /api/company/flywheels, /api/company/failure-modes, /api/company/details, /api/footnotes/revenue-segmentation. FAQ covers economic model vs AI summary, how citation verification works, what happens when no valid quote exists, honest scope limits (verified quote vs perfect interpretation), which SEC filings feed the model, subscription tiers (Stock or Professional plan, plus the one-time 50-request premium sampling allowance), and MCP/AI-agent consumption.
- [Debt Tranche Structure API: SEC Filing Data](https://developer.stockfit.io/blog/debt-tranche-structure-api): How to get a company's debt tranche maturity schedule, per-instrument coupon rates, and credit facility capacity via API, sourced from the XBRL debt-instrument-axis footnote of 10-K and 10-Q filings, not the single total-debt line on the balance sheet. Endpoints covered: /api/footnotes/debt-structure (face amount, carrying amount, stated and effective interest rate, variable-rate spread, fair value, conversion price, and derived due year per tranche, plus a type-level rollup) and /api/footnotes/credit-facilities (revolver and term-loan capacity, amount drawn, remaining headroom, utilization, letters of credit, commitment fees), both on the Professional plan, with /api/financials/balance-sheet as the free-tier on-ramp. Worked end to end on Oracle Corp (ORCL) FY2026 (period 2026-05-31): 56 outstanding tranches totaling $123.5 billion in face value, cross-verified tranche by tranche against Oracle's own SEC prospectus filings for its September 2025 $18 billion and February 2026 $25 billion bond offerings (14 tranches, $43 billion combined, priced to fund the Stargate AI data-center buildout with OpenAI), a fixed-vs-floating-rate breakdown, a maturity-wall chart bucketing face amount by due year, and the $10 billion commercial paper program plus term loan and interest-rate swap from the credit-facilities endpoint. Also covers convertible debt and the conversionPrice dilution field using Wayfair (W), including a live capital-structure change caught between fiscal quarters (a convertible note fully retired, a new non-convertible note issued). Explains the XBRL debt-instrument-axis sourcing via Arelle, data-honesty notes (derived due year, mis-tagged whole-percent rates, US-GAAP-only coverage), and how the /api/footnotes/debt-structure "types" rollup reconciles against balance-sheet total debt. FAQ covers what a debt tranche is, face amount vs carrying amount, detecting fixed vs floating rate debt, credit facility utilization as a liquidity signal, and convertible-debt dilution risk. Not investment advice.
- [Best AI ETFs in 2026: What to Watch This Year](https://developer.stockfit.io/blog/best-ai-etfs-2026): Research-first comparison of five AI-themed ETFs, VanEck Semiconductor ETF (SMH), Global X Artificial Intelligence & Technology ETF (AIQ), Global X Robotics & Artificial Intelligence ETF (BOTZ), iShares Future AI & Tech ETF (ARTY), and Roundhill Generative AI & Technology ETF (CHAT), graded on real exposure to seven independently researched AI-infrastructure stocks (Nebius Group, CoreWeave, Vertiv Holdings, Arista Networks, Marvell Technology, Astera Labs, Credo Technology) rather than a single Nvidia-weight metric, using /api/fund/holdings (full N-PORT holdings, not just top 10) and /api/fund/reverse-lookup (which funds hold each thesis ticker) alongside /api/fund/composition, /api/fund/chart/industry-groups, /api/fund/overlap, /api/fund/fees, and /api/fund/profile. Live figures as of each fund's most recent N-PORT filing (2026-03-31 through 2026-04-30): combined thesis-stock exposure ranges from 0% (BOTZ, an industrial-automation fund holding none of the seven) to 13.95% (ARTY) and 13.62% (CHAT, the only fund holding all seven thesis names); SMH and AIQ, the two largest and most recognized AI ETFs, land under 3% combined despite that popularity. The reverse-lookup method also surfaced a sixth fund outside the original candidate set, BAI (iShares A.I. Innovation and Tech Active ETF), with 4.96% combined thesis exposure (Vertiv, Arista, Credo) and, more notably, direct privately-marked stakes in Anthropic Series G (0.44%) and OpenAI Series C (0.35%) held at fair-value level 3. Also covered as supporting context: Nvidia weight ranges from 2.70% of AIQ to 19.64% of SMH; top-10 concentration from 42.7% (AIQ, 89 holdings) to 72.2% (SMH, 26 holdings); semiconductor vs software-and-SaaS industry-group weight per fund; pairwise holdings overlap, including ARTY and CHAT's substantially redundant 15-holding overlap despite being this post's two thesis-exposure leaders; SEC Form N-1A expense ratios from 0.35% (SMH) to 0.75% (CHAT); and net-assets growth since 2022 ranging from 1.6x (BOTZ) to 67x (AIQ), tracking Nvidia-trade popularity more than real thesis-stock exposure. Also covers multi-year performance: indexed share-price growth per fund (from 397% for SMH since 2021 to 140% for BOTZ since 2016), trailing 1/5/10-year returns from each fund's own N-CSR shareholder report (CHAT's 1-year return went from 7.98% to 122.06%), and N-1A fee-history stability (SMH, AIQ, BOTZ, ARTY, and CHAT's expense ratios effectively unchanged for years; bonus fund BAI's gross rate cut from 0.68% to 0.65% with net held flat at 0.55%), including ARTY's 2024-08-12 rename from IRBO (iShares Robotics and Artificial Intelligence Multisector ETF, a different underlying index) to its current AI mandate, which splits its longer-run price and return figures across two different fund identities. Five charts built from captured, source-cited data: thesis-stock exposure by fund, Nvidia weight by fund, industry-group weight by fund, net-assets growth by fund, and indexed price growth by fund. Endpoints: /api/fund/holdings, /api/fund/reverse-lookup, /api/fund/composition, /api/fund/chart/industry-groups, /api/fund/overlap, /api/fund/fees, /api/fund/profile, /api/fund/exposure-model, /api/fund/performance, /api/price/history. FAQ covers which AI ETF has real AI-infrastructure exposure beyond Nvidia, which has the most Nvidia exposure, how the five funds differ, holdings overlap, fee comparison, N-PORT data freshness, concentration risk, which ETF holds OpenAI or Anthropic stock directly, and how returns/prices/fees have changed over the past few years. Not investment advice.
- [How to Analyze ETF Thematic Exposure via API](https://developer.stockfit.io/blog/etf-thematic-exposure-api): Developer tutorial on analyzing ETF thematic exposure via /api/fund/exposure-model, the audit-grade, lazily-generated fund classification endpoint that reads a fund's prospectus (485BPOS/497/N-1A), N-CSR shareholder reports, N-CEN census, and N-PORT holdings and returns named strategy buckets, each backed by a machine-verified citation (a clickable SEC EDGAR filing URL, the document section, and a verbatim 20-400 character quote confirmed to appear word-for-word in the filing before it is stored). Explains why a plain ETF sector-weighting or asset-exposure API (the kind Financial Modeling Prep, Intrinio, and Massive ship) can only answer "what percentage of assets sits in each sector," a derived statistic with no citation, while the exposure model answers "what does this fund's own prospectus claim its strategy is." Worked end-to-end on ARK Innovation ETF (ARKK): the fund's four named sub-themes (Genomics/"Genomic Revolution Companies," Automation and robotics, Next Generation Internet, Blockchain and Fintech Innovation) each quoted directly from ARKK's 485BPOS prospectus, cross-checked against real portfolio data from /api/fund/composition (top 10 holdings as of 2026-04-30: Tesla, Tempus AI, AMD, CRISPR Therapeutics, Shopify, Roku, Robinhood, Coinbase, Circle, Palantir, 49.5% of the fund) and /api/fund/chart/industry-groups, which shows ARKK's Semiconductors weight rising from under 2% through 2024 to 10.94% by April 2026 as AMD entered the top holdings, charted alongside SPY's steadier semiconductor climb from 3.29% to 14.13% over the same multi-year window. Also covers the endpoint's lazy-generation behavior (a 202 Accepted response with a jobId for a fund whose model has never been requested, resolving to 200 OK within a few minutes) and /api/fund/reverse-lookup for finding every one of the 1,639 funds that hold a given stock like AMD. Endpoints: /api/fund/exposure-model, /api/fund/chart/industry-groups, /api/fund/composition, /api/fund/reverse-lookup. FAQ covers thematic exposure vs sector weighting, the three-call analysis workflow, lazy-generation and job polling, how it differs from FMP's sector-weighting API, self-verifying citations, N-PORT data freshness, required subscription tier (ETF or Professional plan), and finding every fund that holds a stock.
- [Alpha Vantage Alternatives: 5 APIs for 2026](https://developer.stockfit.io/blog/alpha-vantage-alternatives): Vendor-comparison roundup ranking five Alpha Vantage alternatives for SEC fundamentals data in 2026: StockFit, EODHD, Financial Modeling Prep (FMP), Finnhub, and Massive (formerly Polygon.io). Explains why teams outgrow Alpha Vantage: its free tier was tightened from 500 to 25 requests/day in 2024, its fundamentals are normalized to a fixed schema with no raw XBRL fact, no filing date, and no SEC accession number anywhere in the response, and its free tier's terms of service restrict it to personal, non-commercial use. Every competitor fact verified against that vendor's own pricing page, docs, or a live response as of 2026-08-20, including a direct fetch of Alpha Vantage's premium pricing page, documentation, and terms of service, and Massive's post-rebrand pricing at massive.com/pricing. StockFit ranks first: a 300-request/day free tier (12x Alpha Vantage's 25/day) that already includes real audit-grade financial statements via /api/financials/income-statement, /api/financials/balance-sheet, and /api/financials/as-reported, each carrying a filing date and SEC accession number for point-in-time-safe backtesting, illustrated with a live, unedited Costco Q3 FY2026 income-statement response (accession 0000909832-26-000051) and a live /api/company/peers call ranking Walmart, Target, Dollar General, Macy's, BJ's, and Dollar Tree by revenue proximity to Costco, a niche comparable-company endpoint no competitor in the roundup offers. Honest gap disclosure: StockFit's free tier is personal-use only, the same restriction Alpha Vantage and Finnhub apply to theirs, and commercial use starts at the $19/mo Starter plan. A six-provider comparison table (Alpha Vantage included as the baseline row) scores free-tier daily limits, whether fundamentals are included on the free tier, SEC audit-trail presence, cheapest paid fundamentals plan, and data source for all six. Per-vendor verified facts: EODHD's free tier (20/day) excludes fundamentals entirely, with the Fundamentals Data Feed starting at $59.99/mo but covering non-US equities back to 2000 that Alpha Vantage covers thinly; Financial Modeling Prep has the highest free-tier volume (250 requests/day, reconfirmed live) and the broadest general catalog, with annual-only statements on the free tier and its current paid-tier pricing for full quarterly history flagged as unconfirmed since FMP's pricing page renders client-side; Finnhub's free tier (60 req/min) is explicitly personal/non-commercial with fundamentals priced as separate modular add-ons from about $50/month per data category; Massive (Polygon.io's 2026 rebrand) is a market-data-first product whose free Basic plan (5 req/min, end-of-day only) excludes fundamentals, which are a $29/month standalone add-on or bundled into the $199/month Advanced tier. Also covers what to require in an Alpha Vantage replacement (usable free tier, point-in-time correctness, an audit trail, depth beyond the big three statements, honest commercial licensing), a decision guide by use case, and a 6-question FAQ (best Alpha Vantage alternative, best free tier, SEC filings/XBRL availability, commercial use on the free tier, cheapest audit-trail-backed fundamentals, and pricing-model comparison). Links to the deeper /blog/stockfit-vs-alpha-vantage-fundamentals endpoint-by-endpoint teardown and /blog/sector-industry-metrics-vs-fmp. Endpoints referenced: /api/financials/income-statement, /api/financials/balance-sheet, /api/financials/as-reported, /api/company/peers.
- [How to Detect Stock Splits in Historical Data](https://developer.stockfit.io/blog/detect-stock-splits-historical-data): Engineering teardown of programmatic stock-split detection from SEC XBRL, the logic behind /api/company/stock-splits, /api/company/recent-stock-splits, and /api/company/upcoming-stock-splits. Explains why price-jump heuristics fail (splits are invisible once a source split-adjusts historical prices), where split evidence actually lives (us-gaap:StockholdersEquityNoteStockSplitConversionRatio1 and its pre-2015 predecessor, tagged either as a plain fact or only on the class-of-stock dimension), and the core problem: the ratio tag's forward/reverse direction is not trustworthy on its own. Covers the share-count cliff corroboration method (bracketing a candidate split date with us-gaap:CommonStockSharesOutstanding / dei:EntityCommonStockSharesOutstanding observations before and after) and its asymmetric logic: a corroborated share-count drop unambiguously overrides a mistagged direction to reverse, a corroborated rise is never forced to forward (could be an IPO or equity issuance instead), and a magnitude cap above 1,000x excludes pre-IPO recapitalizations. Real, verified split events used throughout: NVIDIA's 10-for-1 (2024) and 4-for-1 (2021), Alphabet's 20-for-1 (2022, disclosed 89 days before the split executed), Amazon's 20-for-1 (2022), Tesla's 3-for-1 (2022) and 5-for-1 (2020), Chipotle's 50-for-1 (2024), vTv Therapeutics' 1-for-40 reverse split mistagged as forward in its own filing and corrected by the cliff, The Honest Company's 2-for-1 split disclosed 559 days after the event, Novavax's 1-for-20 reverse split, Domo's 1-for-15 reverse split, and Ashford Hospitality Trust's three separate 1-for-10 reverse splits (2020, 2021, 2024) compounding to a 1,000x cumulative reversal, a listing-compliance distress pattern charted across the post. Also covers XBRL date-field semantics (why the tagged split date can be a record date, distribution date, or an unrelated quarter-boundary artifact depending on the filer) and how split-adjustment feeds downstream into /api/financials/income-statement's splitAdjust=true default. Two charts built from captured, source-cited data: a 14-event XBRL disclosure-lag chart and Ashford Hospitality Trust's cumulative reverse-split factor. Endpoints: /api/company/stock-splits, /api/company/recent-stock-splits, /api/company/upcoming-stock-splits, /api/financials/income-statement. FAQ covers detection methodology, why price data alone fails, wrong-direction tags, the share-count cliff, disclosure lag, false-positive rises, forward vs reverse splits, ETF/fund split coverage, and subscription tiers.
- [Best Insider Trading APIs in 2026 (7 Compared)](https://developer.stockfit.io/blog/best-insider-trading-api): Vendor-comparison roundup of the seven best insider trading (SEC Form 3/4/5) APIs for developers in 2026: StockFit, Financial Modeling Prep (FMP), Quiver Quantitative, API Ninjas, The Washington Service, SecuritiesDB, and the official free SEC EDGAR API. Each is scored on market-wide screening (query across every ticker vs one company at a time), cluster/aggregate signal detection (distinct-buyer counts, acquired/disposed rollups), SEC accession-number traceability, free tier and commercial-use rights, and related-party coverage. Every competitor claim verified against that vendor's own pricing and documentation as of 2026-08-19; every StockFit claim verified against a live production API response the same day. StockFit ranks first for cluster detection and single-company research: /api/insider-transactions/summary returns uniqueBuyers/uniqueSellers rollups over 3/6/12-month windows (the building block for cluster-buy detection), /api/insider-transactions/roster and /api/insider-transactions/by-insider give a full per-company insider roster and per-person trading history, and every transaction carries its SEC accession number and filing URL, deduplicated across co-filers. Honest gap disclosure: StockFit's insider-transactions family has no free tier (Stock plan $39/mo or Pro $69/mo required) and is scoped per-company with no market-wide "latest insider buys across every ticker" feed, unlike FMP, API Ninjas, and Quiver Quantitative. The post's proof section is built on StockFit's own published empirical study, the insider cluster-buys backtest (45 discretionary clusters, 40 S&P 500 tickers, 2018-2024): a mean +3.1% 365-day excess return vs SPY but a sub-50% hit rate, and the strongest finding, mixed officer+director clusters outperform while single-role clusters do not, illustrated with a bar chart and linked to the full /blog/insider-cluster-buys-sp500 methodology post. Per-vendor verified facts (2026-08-19): Financial Modeling Prep has a market-wide Latest Insider Trading endpoint and quarterly acquired/disposed statistics reachable on its 250-call/day free tier for general data, though insider-trading endpoints are marked a premium dataset with the exact required plan unpublished; Quiver Quantitative has no free tier, Hobbyist ($30/mo or $25/mo annual) excludes insider data, Trader ($75/mo or $62.50/mo annual, $750/yr) includes it, and both plans explicitly exclude commercial-use rights (only the custom-priced Commercial plan grants them); API Ninjas has a real free tier (3,000 calls/month, 100/hour, no commercial use) with accession numbers on every record and ticker-optional market-wide filtering, commercial rights start at $39/mo Developer (100,000 calls/month); The Washington Service is enterprise-only with no public pricing, covering Form 4 and Form 144 plus related-party (trust/spouse/child) resolution no other vendor here documents; SecuritiesDB is a free, no-signup, ticker-scoped buy/sell-ratio endpoint with a short recent-transactions array, not full paginated history; the official SEC EDGAR API (data.sec.gov) is free and keyless, rate-limited to 10 requests/second. Also covers what Form 3/4/5 and Section 16 actually mean, a comparison-criteria section, a decision guide by use case, and a 9-question FAQ (best insider trading API in 2026, best free option, Form 3 vs 4 vs 5, what a cluster buy is and which APIs detect it, market-wide screening support, whether cluster buying predicts outperformance, insider vs beneficial-ownership APIs, data freshness, and accession-number availability). Endpoints referenced: /api/insider-transactions, /api/insider-transactions/summary, /api/insider-transactions/roster, /api/insider-transactions/by-insider.
- [ETF Custodian and Service Provider API](https://developer.stockfit.io/blog/etf-custodian-service-provider-api): Developer guide to looking up an ETF or mutual fund's operational service-provider roster via /api/fund/service-providers (SEC Form N-CEN annual census, grouped by reporting year: custodian, administrator, transfer agent, adviser and sub-adviser, securities-lending agent, pricing service, authorized participant, and broker, each with provider name, LEI, CRD number, isAffiliated flag, and aggregate commission where disclosed), /api/fund/structure (N-PORT plus N-CEN classification flags: isIndexFund, isNonDiversified, isInKindEtf, sector and industry allocation by quarter, geographic exposure, balance-sheet trajectory), and /api/fund/fund-health (fair-value hierarchy, liquidity classification, operational red flags: hasMaterialWeakness, hasLegalProceedings, hasValuationChanges, hasPaymentErrors), with /api/fund/profile as the free-tier fund-identity on-ramp. Explains why holdings and price-focused ETF data vendors cannot answer who custodies a fund's assets, since custodian, transfer agent, and lending agent exist only in the N-CEN operational census, not a holdings file. Live JSON verified 2026-08-17: JEPI's full 44-entity N-CEN roster across 8 roles for its 2025-06-30 filing (22 authorized participants, 11 brokers, 3 custodians, 2 administrators, 1 transfer agent, 1 lending agent, 1 adviser, 3 pricing services); a detected custodian change (BofA Securities, Inc. dropped and Barclays Capital, Inc. added to JEPI's custodian roster between the 2024-06-30 and 2025-06-30 N-CEN filings, Citigroup and JPMorgan Chase unchanged); a second detected custodian change (CIBC Mellon Trust Company added alongside The Bank of New York Mellon on ARKK's 2025-07-31 filing); and JEPI's N-CEN operational history (a hasPaymentErrors flag on the 2021-06-30 filing only, and securities lending authorized from 2020 but not actually used until the 2022-06-30 filing). Covers ETF custodian API, ETF service provider history API, SEC Form N-CEN explained, N-CEN vs N-PORT, authorized participant roster, securities-lending agent, and fund administrator and transfer agent data. Endpoints: /api/fund/service-providers, /api/fund/structure, /api/fund/fund-health, /api/fund/profile. FAQ: how to find an ETF's custodian, what Form N-CEN discloses about service providers, how to detect a custodian change, custodian vs transfer agent, why the authorized-participant roster matters, mutual fund coverage, data freshness, and subscription tiers (ETF and Professional plans).
- [ETF Expense Ratio and Fee API](https://developer.stockfit.io/blog/etf-fee-analysis-api): Developer guide to comparing ETF expense ratios and fees programmatically via /api/fund/fees (SEC Form N-1A prospectus fee table: management fee, 12b-1 fee, other expenses, acquired fund fees and expenses/AFFE, total expense ratio, fee waiver, net expense ratio, paginated across every filed version with accession number and filedDate) and /api/fund/fee-analysis (SEC Form N-CEN operational census: isExpenseLimitation and isExpenseReducedOrWaived flags, aggregate brokerage commissions by counterparty, in-kind creation/redemption percentages, line-of-credit size), with /api/fund/profile as the free-tier fund-identity on-ramp. Explains why consumer comparison tools (ETF Database, Schwab, ETFSift, ETF BFF) show only a single rendered percentage with no API, and why a fee-schedule endpoint sourced from Form N-1A correctly returns empty for unit investment trust ETFs like SPY and QQQ that do not file N-1A the way open-end funds do. Live JSON verified 2026-08-15: PDBC's disclosed gross-to-net gap (0.59% management fee plus 0.15% acquired fund fees and expenses from its Cayman subsidiary equals a 0.74% total expense ratio, offset by a 0.15% waiver back to a 0.59% net expense ratio, filed 2026-02-26, accession 0001104659-26-020374); JEPI's standing-but-unbound expense limitation agreement (N-CEN flags isExpenseLimitation and isExpenseReducedOrWaived true every period since 2020, N-1A fee table flat at 0.35% with no populated waiver since 2022); a six-fund net-expense-ratio comparison (SCHD 0.06%, JPST 0.18%, JEPI 0.35%, MOAT 0.46%, PDBC 0.59%, ARKK 0.75%); and a JEPI-vs-ARKK net-assets trajectory chart from SEC Form N-PORT (JEPI growing from $0.17B in 2020 to $43.96B in 2026, ARKK falling from a $21.36B 2021 peak to about $6.48B), with napkin-math implied annual fee dollars illustrating why a lower sticker expense ratio on a growing fund can cost more in real terms than a higher one on a shrinking fund. Covers ETF expense ratio comparison, ETF fee analysis API, true cost of an ETF beyond the expense ratio, gross vs net expense ratio, acquired fund fees and expenses (AFFE), and N-CEN expense limitation agreements. Endpoints: /api/fund/fees, /api/fund/fee-analysis, /api/fund/profile, /api/fund/fund-health. FAQ: how to compare ETF expense ratios and fees via API, gross vs net expense ratio, why SPY/QQQ return no fee-schedule results, what an expense limitation agreement means, what AFFE is, data freshness, and subscription tiers (ETF and Professional plans).
- [SEC S-1 IPO Data API: Filing to Effective Date](https://developer.stockfit.io/blog/sec-s1-ipo-data-api): Developer guide to the full IPO filing lifecycle on SEC EDGAR: Form S-1 registration statements, S-1/A price-range amendments, Form 424B4 final prospectuses, and the standalone EFFECT filing that is the only authoritative signal a registration statement has become effective and the stock is cleared to trade, distinct from any price-range amendment or press release. Walks Figma's (FIG) real July 2025 IPO end to end via the StockFit /api/filings/offering endpoint: the initial S-1 filed 2025-07-01 with no price, the S-1/A filed 2025-07-21 at a $25 to $28 range (12,472,657 shares, 5,540,561 overallotment), the S-1/A filed 2025-07-28 raising the range to $30 to $32, and the EFFECT filing dated 2025-07-30 with periodEndDate 2025-07-31 matching Figma's actual NYSE trading debut (final price $33, above every disclosed range). Covers the gotcha that a company can file more than one EFFECT record over its life (Circle Internet Group's second EFFECT landed two months after its actual IPO for an unrelated registration), so pipelines must take the first EFFECT per CIK, not the latest. Shows live JSON from /api/filings/recent-offerings (the isPriced flag, small-cap-heavy daily S-1 flow) and /api/filings/stats, plus a TypeScript walkthrough for building an IPO calendar by joining the offering feed to per-CIK EFFECT lookups. Endpoints: /api/filings/offering, /api/filings/recent-offerings, /api/filings/stats, /api/filings (type=EFFECT). FAQ: S-1 vs 424B4, what the EFFECT filing means, how S-1/A price ranges move, handling multiple EFFECT filings per company, S-1 vs S-11 vs F-1, whether the S-1/A range is the final IPO price, and free-tier vs Stock/Pro tier coverage.
- [How to Find All ETFs That Hold a Specific Stock](https://developer.stockfit.io/blog/find-etfs-holding-stock): Developer tutorial on finding every ETF and mutual fund that holds a specific stock programmatically with the StockFit ETF reverse lookup API (/api/fund/reverse-lookup), instead of paste-a-ticker web tools like the etf.com, ETF Database, or TipRanks stock exposure pages that return no JSON and cannot feed a screener, script, or AI agent. One call per stock returns the paginated fund list sorted by position market value, each row carrying fundName, fundTicker, fundCik, fundSeriesId, valueUsd (market value), pctVal (portfolio weight as percent of fund net assets), balance (shares held), reportDate, and fundNetAssets, sourced from SEC Form N-PORT position-level filings parsed from EDGAR XML with no third-party aggregators. Live examples captured 2026-07-24: NVIDIA (NVDA) held by 2,215 funds with Vanguard Total Stock Market (VTI) the largest position at $127.8B, and the three S&P 500 trackers (VOO 7.577%, IVV 7.564%, SPY 7.578%) independently agreeing on NVIDIA's portfolio weight within 1.4 basis points across separate N-PORT filings; Apple (AAPL) held by 1,996 funds; Micron (MU) held by 1,369 funds with a 2.15% QQQ weight. Includes cURL and official Node.js SDK (@stockfit/api, createClient plus fundReverseLookup) code, querying by CUSIP (67066G104), composite FIGI (BBG000BBJQV0), or CIK instead of ticker, walking the full paginated list (pageSize up to 1000; all 2,215 NVDA-holding funds in three requests, $1.20T aggregate NVDA position across registered funds), a watchlist ETF exposure screen loop, the fund-of-funds direction (which funds hold an ETF: QQQ itself held by 124 funds, led by Direxion Monthly NASDAQ 100 Bull 1.75X at a 65.5% weight, with look-through exposure math), reading portfolio weight vs shares held vs market value, N-PORT quarterly freshness vs the daily ETF holdings endpoint, why some mutual fund rows have a null ticker (share classes without tickers in SEC data, identified by CIK and series id), and the fund overlap and holdings endpoints as follow-ups. Covers ETF stock exposure, ETFs that hold a stock, reverse ETF lookup, crowding and concentration screens, and passive flow passthrough. Endpoints: /api/fund/reverse-lookup, /api/fund/holdings, /api/fund/holdings/daily, /api/fund/overlap. FAQ: how to find all ETFs holding a stock, what an ETF reverse lookup API is, where ETF holdings data comes from, N-PORT freshness, mutual fund coverage, null tickers, tiers (ETF and Professional plans).
- [Best Fundamentals APIs in 2026 (8 APIs Compared)](https://developer.stockfit.io/blog/best-fundamentals-api): A vendor-comparison roundup of the eight best fundamentals APIs (financial data / financial statements APIs) for developers in 2026: StockFit, Financial Modeling Prep (FMP), EODHD, Alpha Vantage, Finnhub, Intrinio, Financial Datasets AI, and sec-api.io. Each is scored on free tier, cheapest paid fundamentals plan, data source (SEC XBRL vs third-party normalized), coverage and historical depth, whether returned values carry a point-in-time SEC accession audit trail, and MCP support, with every competitor claim verified against that vendor's own pricing and documentation as of 2026-07-14. Evaluation criteria covered: data source and traceability, as-reported vs normalized financial statements, point-in-time correctness for backtesting, free tier, coverage and historical depth (quarterly and annual income statement, balance sheet, cash flow statement), financial ratios and valuation metrics, developer experience (REST API, structured JSON, rate limits, MCP server), and scope honesty (real-time market data, OHLCV and intraday data, technical indicators, and options chains are out of scope for a pure fundamentals API). StockFit ranks first for audit-grade, SEC-traceable fundamentals: it parses XBRL from 10-K, 10-Q, and 20-F filings via Arelle, returns both as-reported and normalized statements, and attaches the SEC accession number and filing date to every value with an amendment trail (point-in-time), on a free tier of 300 calls per day; illustrated with a six-year Apple revenue chart where each bar traces to its 10-K accession number (FY2025 revenue 416.161 billion dollars, accession 0000320193-25-000079) and a real /api/financials/income-statement JSON response showing the sources map keyed by accession. Per-vendor verified facts (2026-07-14): FMP free 250/day is prices not fundamentals, cheapest fundamentals about $22/mo annual (annual-period only) or about $59/mo Premium (quarterly plus 30 years), self-standardized from SEC filings, MCP yes; EODHD free 20/day excludes fundamentals, Fundamentals Data Feed $59.99/mo ($49.99 billed annually), third-party normalized across 70-plus exchanges, MCP yes; Alpha Vantage free 25/day includes full normalized statements, premium from $49.99/mo ($41.58 annual) buys only a higher rate limit not more data, MCP yes; Finnhub free 60/min includes SEC as-reported statements plus key metrics but the normalized statements are the $50/mo Fundamental Data add-on, no official MCP; Intrinio has no production free tier (Dow-30 sandbox), Individual $150/mo, SEC XBRL standardized plus as-reported, roughly 10-30 minutes post-filing, MCP yes; Financial Datasets AI free for five tickers (AAPL, MSFT, NVDA, TSLA, GOOGL), $20 one-time credits (one year of history) or Build $200/mo (30-plus years), SEC-sourced normalized plus as-reported with a per-filing accession_number and filing_url, MCP yes; sec-api.io free is a one-time 100-call trial, Personal and Startups $55/mo ($49 annual), raw XBRL-to-JSON as-reported statements with no pre-computed ratios (you derive them yourself), no MCP. Also covers a decision guide (which API to pick by use case), honorable mentions (Twelve Data, Tiingo, Polygon now Massive) and the free official SEC EDGAR API baseline (10 requests per second), and an 8-question FAQ (best fundamentals API in 2026, best free fundamentals API, most accurate fundamentals API, what an SEC accession number is and why it matters, as-reported vs normalized fundamentals, best for AI agents, best for backtesting, quarterly and historical coverage). Endpoints referenced: /api/financials/income-statement, /api/financials/as-reported, /api/financials/key-metrics, /api/filings.
- [Accurate Earnings Dates API: Predict and Confirm](https://developer.stockfit.io/blog/accurate-earnings-dates-api): How StockFit produces accurate earnings dates with a two-layer, predict-then-confirm approach, and how to tell a confirmed date from a predicted one in the API response. Layer one predicts each company's next earnings date from its own SEC filing cadence on EDGAR (when the same-quarter 10-Q or 10-K was filed in prior years, adjusted for fiscal-calendar drift), which for large- and mid-cap issuers lands within a few business days; this is what /api/earnings/date returns for a far-out name like NVIDIA (projected 2026-08-26 with no report time). Layer two, inside the near-term window of roughly the next 28 days (the look-ahead ceiling on /api/earnings/upcoming), cross-checks and firms up the date against official exchange data (Nasdaq-based, not press-release scraping or a black-box third-party algorithm), adding the expected report time (pre_market or after_hours) and the analyst consensus EPS estimate. The machine-readable line between the two layers is the reportTime field: near-term confirmed dates carry it (JPMorgan 2026-07-14 pre_market, Alphabet 2026-07-22 after_hours, Microsoft 2026-07-29, Apple 2026-07-30), far-out predicted dates omit it (NVIDIA). Fields are omitted rather than nulled, so a simple presence check tells you which layer a date came from. Exchange-sourced provenance is the trust factor most calendars are quiet about: a date maintained by the venue where the security trades carries more weight than one a scraper inferred from an investor-relations page, and it fits the broader StockFit model where every fundamental value carries its SEC accession number. The absolute source of record is the company's own 8-K filed under Item 2.02 (Results of Operations and Financial Condition, press release attached as Exhibit 99.1), streamed market-wide via /api/filings/latest with event=earnings (each record carrying the accession number and EDGAR url), with the body extractable via /api/filings/item item=2.02. The three layers stack: predict far out, confirm near term, record on filing. Distinct from the earnings calendar API developer guide (which is the full endpoint walkthrough); this post is specifically about date accuracy and provenance. Endpoints: /api/earnings/date, /api/earnings/calendar, /api/earnings/upcoming, /api/filings/latest, /api/filings/item, /api/earnings/eps-history, /api/earnings/dividend-history. Includes live production JSON examples (as of 2026-07-12), the estimated-vs-confirmed comparison, tier coverage (Starter for single-ticker date, Stock and Pro for market-wide and batch), and an 8-question FAQ on how accurate earnings dates are, estimated vs confirmed, where the dates come from, and how far in advance they are confirmed.
- [Daily ETF Holdings: Six File Formats, One API](https://developer.stockfit.io/blog/daily-etf-holdings): Engineering build-in-public story on the StockFit daily ETF holdings API (/api/fund/holdings/daily), which returns a fund's full current portfolio sourced from the issuer's own website once a day, far fresher than the quarterly SEC N-PORT filing behind /api/fund/holdings. Explains why daily holdings beat quarterly N-PORT: registered funds file Form N-PORT monthly but only the quarter-end month is disclosed publicly, about 60 days after the quarter closes, so a public holdings list is one to four months stale, while the issuer's website file (mandated by the SEC ETF Rule, Rule 6c-11, since 2019) carries the prior close. The core engineering problem is format chaos: there is no standard, so issuers publish the daily file in six different formats, and one API normalizes all of them. Verified production coverage as of 2026-08-16: 4,902 US-listed funds across 391 distinct issuer sites (more than four out of five US-listed ETFs), distributed by source publish format as csv 1,963 (40.0%), html tables 1,098 (22.4%), json 1,018 (20.8%), xlsx OOXML workbooks 591 (12.1%), xls legacy binary BIFF 175 (3.6%), and pdf 57 (1.2%). Coverage by issuer (top families): iShares/BlackRock 435, First Trust 313, Invesco 228, SPDR/State Street 178, ProShares 166, Innovator 155, Direxion 133, Vanguard 115, Global X 113, WisdomTree 94, Defiance 88, VanEck 83; onboarding one issuer's format unlocks its whole lineup, and the supported-funds list grows as more issuers are added (the authoritative roster is /api/fund/holdings/daily/supported-funds). The endpoint also keeps point-in-time history: every day's snapshot is stored with roughly 30 days of rolling retention per fund (the most recent snapshot always kept, accumulating since early August 2026), and passing reportDate=YYYY-MM-DD returns the most recent snapshot on or before that date, with the actual as-of day served in the response's reportDate field (a minority of issuers publish only month-end or quarter-end holdings). Funds are onboarded by writing a config (source URL, format, field mapping), not by shipping per-fund code, and the hard parts are access tricks (browser user-agent files, monthly-rotating URL tokens discovered via a documents API, TLS-fingerprint 403s that are false bot walls, and JavaScript challenges cleared by a real Chromium navigation). The unified JSON response carries reportDate (the issuer's as-of date), name, ticker, cusip, balance, units, currency, valueUsd, and pctVal, sorted by portfolio weight descending, with only the fields the source provides. Live examples: ARK Innovation ETF (ARKK, a csv source, as-of 2026-08-14, TSLA 9.16%, SpaceX 5.98%, TEM 5.21%), the senior-loan fund SRLN (as-of 2026-08-13, 690 holdings) where the schema adapts to fixed income with units PA (par), couponRate, and a normalized maturityDate in YYYY-MM-DD, and a point-in-time query (CNEQ with reportDate=2026-08-10 returning that day's 32-position snapshot). Honest scope limits: the daily feed covers transparent US-listed ETFs only, not semi-transparent or non-transparent active ETFs (proxy basket daily, full book quarterly), not mutual funds or closed-end funds (quarterly N-PORT, monthly workbook at best), and the daily rows are not mapped to a CIK or mappedSymbol and carry no fair-value level. Endpoints: /api/fund/holdings/daily, /api/fund/holdings/daily/supported-funds, /api/fund/holdings, /api/fund/overlap. Includes two Highcharts visualizations (format distribution, coverage by issuer), curl and JSON examples for an equity fund, a bond fund, and a point-in-time history query, and a 9-question FAQ covering daily-vs-N-PORT freshness, file formats, historical snapshots, bond fields, and which funds are supported.
- [Executive Compensation API: Pay vs Performance](https://developer.stockfit.io/blog/executive-compensation-api): Data-story walkthrough of the StockFit executive compensation API (/api/executives/compensation), which returns CEO and Named Executive Officer pay parsed from the SEC-mandated pay-versus-performance table in DEF 14A proxy statements (Item 402 of Regulation S-K) as clean JSON, with the reported Summary Compensation Table total, Compensation Actually Paid (CAP, called Actually Paid Compensation in the post), the average for non-CEO Named Executive Officers, and company and peer-group Total Shareholder Return indexed to $100, each row traceable to its SEC accession number. Explains the two pay numbers: the Summary Compensation Table reports equity at grant-date fair value (FASB ASC Topic 718, the accounting cost on grant day), while Actually Paid Compensation (the 2022 Item 402(v) pay-versus-performance rule) re-marks equity to its actual change in value during the year, so it moves with the share price and can go negative in a down year. Worked example on NVIDIA (Jensen Huang) across FY2021-FY2026: reported total stayed roughly $19M to $50M while Actually Paid swung from negative $4.1M (FY2023, stock fell) to $344.2M (FY2025, company TSR 1,100 vs peer 168), traced to DEF 14A accession 0001045810-25-000095. Extreme single-year CEO cases: Broadcom Hock Tan FY2024 reported $2.6M but actually paid $1.15B, Oracle Safra Catz FY2025 reported $1.1M ($950k salary, no bonus, no new grant) but actually paid $462M, Chevron and Mondelez CEOs posted negative Actually Paid in a down equity year. Cross-sector teardown of 30 large caps in six sectors (Technology NVDA MSFT AAPL ORCL AVGO, Financials JPM GS MS BAC WFC, Energy XOM CVX COP SLB EOG, Healthcare JNJ UNH MRK ABBV ABT, Consumer Staples KO PG PEP MDLZ CL, Utilities NEE DUK SO AEP D): FY2024 Actually Paid as a multiple of reported NEO pay runs about 4.9x for technology and 2.4x for financials (equity-linked pay swings with the market), versus roughly 1.0x to 1.3x for utilities, healthcare, and staples, and below 1.0x for energy (paid less than reported as energy shares lagged). Also covers the company-selected performance measure (NVDA non-GAAP operating income, JPM return on tangible common equity, XOM cash flow from operations and asset sales, KO organic revenue growth) via /api/executives/performance-measures, and compensation-governance flags (insider-trading policy, predetermined award timing, MNPI considerations) via /api/executives/governance. Closes on the free-tier on-ramp (/api/filings and /api/financials/as-reported) and notes the CEO pay ratio (Item 402(u)) is not exposed. Endpoints: /api/executives/compensation, /api/executives/performance-measures, /api/executives/governance, /api/executives/officers, /api/filings, /api/filings/search-by-accession-number, /api/financials/as-reported. Includes three Highcharts visualizations (sector divergence bars, NVIDIA reported-vs-actually-paid columns with a negative year, per-company log-scale ratios), two data tables, curl and JSON examples, and a 7-question FAQ.
- [SEC EDGAR Alternative Data for Quants](https://developer.stockfit.io/blog/sec-edgar-alternative-data-signals): A product-led list of the eight most unconventional, hard-to-extract datasets StockFit pulls out of SEC EDGAR for quantitative research and backtesting, and the signal each one uncovers, framed for small quant pods and emerging managers priced out of conventional alternative data (credit-card panels, satellite, Bloomberg). The thesis: everyone can download EDGAR for free, but the edge is extraction depth, and most vendors stop at the normalized income statement while StockFit goes into the dimensional XBRL axis, the proxy pay tables, the ownership amendment trail, and N-PORT internals, with every value carrying its SEC accession number for a point-in-time, lookahead-free audit trail. The eight datasets with live verified JSON examples: (1) the audit-grade AI Economic Model (/api/company/economic-model) returning a structured business model (offerings, monetization, cost structure, operating levers, structural advantages, flywheels, capital allocation, failure modes) where every claim carries a clickable EDGAR URL, the document section, and a verbatim quote machine-verified against the filing text at generation time, demonstrated on Apple; (2) business segment operating income (/api/footnotes/business-segmentation) from the XBRL business-segments axis, showing Amazon fiscal 2025 where AWS is about 18% of segment revenue but 57% of segment operating income and absorbed about 96.5 billion dollars of capex, the real profit engine and AI capex cycle; (3) revenue segmentation (/api/footnotes/revenue-segmentation) splitting revenue by geography (countries, US states, regions, residuals) and product, shown on Apple (iPhone 209.6B, Services 109.2B, China 64.4B, US 151.8B for fiscal 2025); (4) Compensation Actually Paid (/api/executives/compensation and /api/executives/performance-measures) from the 2022 SEC pay-versus-performance rule, where Apple reported 74.6M dollars CEO total compensation in fiscal 2024 but CAP was 169.0M, plus the company-selected performance measure; (5) the Schedule 13D vs 13G ownership signal (/api/ownership/beneficial-owners/history) where the filingType field encodes intent and the history endpoint catches the 13G-to-13D passive-to-activist switch, shown on Southwest Airlines (LUV) with activist Elliott on a 13D among passive 13G holders Primecap and Franklin; (6) Form 4 insider transaction codes (/api/insider-transactions) where P is an open-market purchase, S a sale, M an option exercise, F tax withholding, A a grant, G a gift, demonstrated on Apple whose recent activity is all F/S/M/G and zero P (no open-market conviction buying), and the cluster-buy study link; (7) N-PORT fund flows and portfolio overlap (/api/fund/flows and /api/fund/overlap) using filing-sourced sales and redemptions rather than modeled estimates and CUSIP-matched holdings, shown on QQQ vs VGT where 37 shared names are about 49% of QQQ and 71% of VGT (a crowding signal); and (8) sector-aware Altman Z-Score and Piotroski F-Score (/api/financials/scores) computed from point-in-time XBRL, shown on Apple (Piotroski 8 of 9, Altman Z 2.42 grey zone, with the book-value-of-equity input caveat explained). Closes on wiring signals into a lookahead-free backtest by keying every value to its filing date, the free-tier on-ramp (/api/filings and /api/financials/as-reported), and a not-financial-advice note. Endpoints: /api/company/economic-model, /api/footnotes/business-segmentation, /api/footnotes/revenue-segmentation, /api/executives/compensation, /api/executives/performance-measures, /api/ownership/beneficial-owners/history, /api/insider-transactions, /api/insider-transactions/summary, /api/fund/flows, /api/fund/overlap, /api/financials/scores, /api/filings. Includes eight live JSON examples, a segment revenue-vs-operating-income mix visualization, and an 8-question FAQ.
- [Beneficial Ownership API: Schedule 13D and 13G](https://developer.stockfit.io/blog/beneficial-ownership-13d-13g-api): Product walkthrough of the StockFit beneficial ownership API, which returns every Schedule 13D and Schedule 13G filer for a US-listed stock as clean JSON, one call per ticker, parsed straight from SEC EDGAR. Explains the regime: Sections 13(d) and 13(g) of the Securities Exchange Act require disclosure when an investor acquires beneficial ownership of more than 5% of a voting class, and the form encodes intent: control-intent investors (activists, bidders, proxy contests) file Schedule 13D, while passive holders (index funds, ETFs, qualified institutional investors with no control intent) file Schedule 13G. The /api/ownership/beneficial-owners endpoint returns the latest filing per reporting person with reportingPersonName, filingType (13D vs 13G, the authoritative passive-or-activist signal), percentOfClass, aggregateAmountOwned, typeOfReportingPerson (PN partnership, HC holding company, IA investment adviser, IN individual), securityClassTitle, reportDate, accessionNumber, and filingUrl, so every stake is traceable to the exact SEC document (the audit-trail wedge a normalized third-party ownership table cannot match). The /api/ownership/beneficial-owners/history endpoint returns every filing per holder over time (not collapsed to the latest), which is how you detect the high-value 13G-to-13D activist switch: a holder whose Schedule 13G is followed by a later Schedule 13D has flipped from passive to activist, an event that routinely moves the stock. Live verified example on Match Group (MTCH, owner of Tinder and Hinge): activist Starboard Value (Jeffrey Smith) holds via Schedule 13D at 4.6% (accession 0000902664-26-002298, filed 2026-05-06, XBRL-tagged) alongside passive Schedule 13G holders BlackRock 11.9%, Vanguard 6.37%, Ameriprise 5.7%, and State Street 3.6%, rendered as an on-brand Highcharts column chart colored by Schedule type. Covers why the raw EDGAR feed is hard (CUSIP-keyed cover pages, historically free-text exhibits, deep amendment chains with more than 120 13D/13G filings for MTCH since 2017) and the 2023 SEC Modernization of Beneficial Ownership Reporting (Release 33-11180, effective 2024) that shortened the initial 13D deadline from 10 calendar days to 5 business days, set 13D amendments at 2 business days, and mandated structured machine-readable filings. Contrasts the three SEC ownership families and when to use each: beneficial ownership (13D/13G, concentrated 5% strategic stakes with intent), institutional ownership (Form 13F-HR, the wide quarterly institutional book lagging up to 45 days), and insider ownership (Forms 3/4/5, Section 16 officer and director trades). Closes on point-in-time ownership without lookahead bias (use the filing date, not the acquisition date) and tier gating (beneficial ownership on Stock and Professional plans, the raw 13D/13G filing trail on /api/filings free). Endpoints: /api/ownership/beneficial-owners, /api/ownership/beneficial-owners/history, /api/ownership/summary, /api/filings, /api/filings/search-by-accession-number. Includes curl and JSON examples, one Highcharts visualization, a 13D-vs-13G comparison table, and a 7-question FAQ.
- [Point-in-Time Data: Essential for Backtesting](https://developer.stockfit.io/blog/point-in-time-data-backtesting): Why quantitative backtests need point-in-time (PIT), timestamped, normalized SEC fundamentals, and the three ways non-PIT fundamental data silently corrupts a backtest. (1) Lookahead bias from restated financials: most databases overwrite the originally-filed value with the later corrected one, so a simulation set before the restatement uses a number that did not exist yet. Worked example with live StockFit data on Plug Power (PLUG): fiscal-2018 net loss to common stockholders was originally filed at about $78.1M (diluted EPS -$0.36) in the March 2019 10-K, then restated to about $85.7M (diluted EPS -$0.39) in a filing dated 2021-05-14, roughly two years later. Shows the real /api/financials/income-statement JSON audit trail, where each period carries a top-level dateFiled (always the original 10-K/10-Q/20-F/40-F filing date) and a sources map keyed by SEC accession number, and a restating filing records a before delta holding the prior value, so you can reconstruct what was knowable on any past date. (2) Reporting lag: the original number does not exist on the period-end date either; it arrives weeks later when the 10-K is filed. Live chart of nine mega-caps (NVDA, MSFT, AAPL, AMZN, COST, WMT, JPM, HD, XOM) shows a 25 to 49 day gap (mean about 38 days) between fiscal year-end and the 10-K hitting EDGAR, against the SEC large-accelerated-filer deadlines of 60 days for a 10-K and 40 for a 10-Q. (3) Survivorship bias: a universe of only currently-trading tickers drops every bankrupt or delisted name; StockFit retains 2,190 fully-delisted companies (no active listing on any exchange, including OTC) and exposes them on /api/company/delisted, with WeWork (ticker WE, dark in 2023) as a worked example. Also covers normalization (concept proliferation in raw XBRL, fiscal-calendar drift across non-December year-ends, and the missing fourth quarter reconstructed from FY minus 9M), the as-reported vs normalized split (/api/financials/as-reported returns the raw filed fact tree; /api/financials/income-statement maps to a stable canonical schema), and a three-rule recipe for a reproducible PIT snapshot (never use a fact before its dateFiled, never use a restated value before the restatement's filing date, include delisted names). Endpoints: /api/financials/income-statement, /api/financials/as-reported, /api/company/delisted. Includes two Highcharts data visualizations and a 6-question FAQ.
- [Mapping CIK to Fund Tickers: A Practical Guide](https://developer.stockfit.io/blog/map-cik-to-fund-ticker): Why resolving a registered fund's current ticker from its SEC EDGAR CIK is far harder than for an operating company, and the classification framework that gets it right. For stocks, company_tickers.json maps CIK to ticker in one file, but for funds every SEC source fails. The EDGAR submissions feed (submissions.zip, one JSON per CIK) carries tickers only at the entity/trust level, so a multi-series fund trust like ETF Series Solutions or iShares Trust has an empty tickers array even though it sponsors dozens of live ETFs, because tickers live on the series, not the trust. And company_tickers_mf.json, SEC's dedicated mutual-fund and ETF ticker file, carries the stale registration ticker and lags the market by months: renames like SPLG to SPYM stay stale, liquidated funds stay listed, and recycled symbols point at the dead fund (IBIT was a liquidated Defiance ETF before BlackRock reused it for the iShares Bitcoin Trust). Explains why this is structural rather than a bug: one CIK maps to many series and tickers, funds rename, liquidate, and merge constantly, tickers get recycled across unrelated issuers, and SEC's files are registration artifacts, not a live securities master. Presents the framework: (1) classify the filer by what it files, not its name, where 10-K/8-K operating companies and grantor trusts like GLD/IBIT use the entity-level submissions ticker while N-CEN/NPORT-P/497 registered 1940-Act funds derive the ticker from filings; and (2) treat a fund's ticker as an event stream, not a field, where the N-CEN annual census gives the current per-series ticker snapshot, NPORT-P monthly portfolio reports confirm liveness or a final filing, and 497 prospectus supplements carry renames, ticker swaps, liquidations, and mergers, reconciled into one current answer. Covers fund delisting and liquidation detection and survivorship bias. Served via /api/lookup/symbol, /api/lookup/cik, /api/fund/profile, /api/fund/changes, and /api/company/delisted. Includes a 7-question FAQ.
- [Extract EPS and Share Count with Arelle](https://developer.stockfit.io/blog/extract-eps-share-count-arelle): How to extract earnings per share (EPS) and weighted-average shares outstanding from SEC XBRL filings with the Arelle processor, and why roughly 1,100 multi-class filers break a naive extractor. For most companies EPS (us-gaap:EarningsPerShareBasic / EarningsPerShareDiluted) and weighted-average shares (us-gaap:WeightedAverageNumberOfSharesOutstandingBasic / Diluted) are tagged as plain dimensionless XBRL facts that can be read directly. Multi-class issuers like Visa (V), Alphabet (GOOGL/GOOG), Berkshire Hathaway (BRK.A/BRK.B), Fox, and News Corp instead tag EPS and share count once per share class on the us-gaap:StatementClassOfStockAxis dimension under the ASC 260 two-class method, with no consolidated company-wide total, so a parser that keeps only the dimensionless fact returns nothing for them. Explains why you cannot simply sum the per-class share counts (treats a Class C share as equal to a Class A share when it is not) or average the per-class EPS (averaging ratios on different denominators). Presents the as-converted formula: as_converted_shares = sum over classes of shares_class x EPS_class / EPS_primary, then consolidated_EPS = net_income / as_converted_shares, which reconciles to the primary (listed) class's reported EPS. The per-class EPS ratio is itself the as-converted conversion weight. Worked example with real Visa FY2025 numbers: Class A 1,714M shares at $10.22, Class C 9M at $40.87 (weight 4.00x), Class B-2 120M at $15.72, Class B-1 5M at $15.97, summing to 1,942.4M as-converted basic shares and a consolidated basic EPS of $10.22 (diluted 2,194.5M shares at $10.20). Covers the hard parts: picking the primary class, excluding non-participating members (preferred/treasury/warrant/escrow), classes with no per-class EPS, basic vs diluted, and FY/quarter/TTM period handling. Served via /api/earnings/eps-history, /api/earnings/snapshot, and /api/financials/income-statement, with curl and JSON examples and a 7-question FAQ.
- [Revenue Segmentation API: Geography and Product](https://developer.stockfit.io/blog/revenue-segmentation-api): Developer guide to the StockFit Revenue Segmentation API (/api/footnotes/revenue-segmentation), which returns a company's revenue split by geography and by product, period by period, in one call per ticker, parsed from XBRL dimensional facts in 10-K, 10-Q, and 20-F filings. Geography is partitioned into four buckets: countries (ISO 3166-1 alpha-2 leaves with continent), usStates (XBRL stpr:XX postal codes for state-level disclosure), regions (filer-tagged rollups like srt:AmericasMember or us-gaap:EMEAMember, each carrying explainedByCountries and other so country leaves reconcile against their parent region without double-counting and continent attribution is applied only to true top-level rollups), and residuals (filer-extension catch-alls like aapl:OtherCountriesMember that are surfaced, never dropped). The product field is a flat list of revenue-by-product members with parent/leaf subtotals preserved rather than pruned. Explains the data pipeline: SEC EDGAR XBRL dimensional facts processed through the Arelle engine, single-axis revenue facts only (qualifier axes like consolidation-items excluded), period-classified into FY/Q1-Q4/H1/9M against each company's real fiscal-year end, written to a dimensional_facts table where restatements insert as new rows and the newest-filed value wins (point-in-time, original dateFiled preserved). Covers quarterly revenue segmentation with Q2/Q3 synthesized from cumulative H1/9M (Q2 = H1 - Q1, Q3 = 9M - H1). Includes live FY2025 examples: AAPL (US $151.8B, China $64.4B countries plus an Other Countries residual $200.0B; products iPhone $209.6B, Services $109.2B, Mac, iPad, Wearables, plus a Products rollup $307.0B), MSFT and KO (US leaf plus a Non-US region with continent null and other = full total), TSLA (Automotive Revenues parent = Sales + Leasing + Regulatory Credits), and NKE (product-only, empty geography). Contrasts with the sibling business-segmentation endpoint (/api/footnotes/business-segmentation, up to eight per-segment metrics: revenue, operating income, gross profit, assets, goodwill, D&A, capex, net income; MCK example). Includes curl, JSON, and JavaScript examples, Starter/Stock/Professional tier gating, MCP availability, and a 7-question FAQ.
- [Company Peers API: Accurate Comparable Companies](https://developer.stockfit.io/blog/company-peers-api): Guide to the StockFit Company Peers API (/api/company/peers), which returns up to six comparable public companies for a stock, ranked by how close they are in annual revenue, each annotated with its revenue and reporting currency, plus the subject company's own revenue, in one call per ticker. Establishes that a peers endpoint is rare among financial-data APIs: Alpha Vantage and Twelve Data have no peers/competitors/comparable-companies endpoint at all, and Finnhub's /stock/peers returns only a flat array of ticker strings grouped by broad country and sector/industry with no revenue, ranking, size, or currency, and is frequently short or empty. Explains why naive peer lists are wrong: raw SEC SIC codes and broad sector buckets lump infrastructure software with cybersecurity, payments, and video games, and put health insurers (UnitedHealth) next to property-and-casualty carriers. Introduces SFIC (StockFit Industry Classification), StockFit's proprietary in-house alternative to the licensed GICS standard: a curated two-level sector-and-industry taxonomy built on SEC data that splits broad buckets (cybersecurity vs software, managed care vs P&C insurance, steel vs mining), reroutes misfiled mega-caps, and resolves issuers canonically so share-class and listing duplicates do not pollute the list. Peers are ranked by revenue proximity to the subject; non-USD reporters (CAD, CNY) are labeled and sorted last, with USD preferred when a foreign issuer dual-reports. Includes live MSFT, NVDA, UNH (managed-care cohort), and UNP (US rails then CAD-tagged Canadian rails) responses, curl and JavaScript examples, Stock and Professional tier gating, and a comparison table versus Alpha Vantage, Finnhub, and Twelve Data. Includes an 8-question FAQ.
- [Executive Officers API: Company Leadership Data](https://developer.stockfit.io/blog/executive-officers-api): Walkthrough of the StockFit Executive Officers API (/api/executives/officers), which returns a US-listed company's current executive officers and CEO with name, SEC-disclosed title, and firstSeen/lastSeen filing dates. Explains why the obvious source (DEF 14A proxy ecd:PeoName XBRL tag) fails (Microsoft and Alphabet return nothing, the CEO is often unnamed) and how the roster is instead reconstructed from SEC Form 3, 4, and 5 (Section 16) insider filings: deduped by the reporting person's own CIK so spelling and surname changes collapse to one record, names normalized from "Last First" to "First Last", and firstSeen/lastSeen taken from filing dates because Form 3 has no transaction date. Covers the role field (ceo, ceo-divisional, officer) with live Amazon (Andrew Jassy ceo; Matthew Garman and Douglas Herrington ceo-divisional for AWS and Stores) and Netflix co-CEO examples, the 18-month activity-based recency window that drops departed executives, the foreign-private-issuer gap (Form 20-F filers like ASML and SAP are Section 16 exempt and return empty), curl and JavaScript examples, Stock and Professional tier gating, and a note that the broader Executives API also exposes compensation and pay-versus-performance (/api/executives/compensation), governance flags (/api/executives/governance), and company-selected performance measures (/api/executives/performance-measures). Includes a 7-question FAQ.
- [ChatGPT App for Stock Fundamentals and SEC Data](https://developer.stockfit.io/blog/stockfit-chatgpt-app): Launch announcement for the StockFit app on ChatGPT, built on the ChatGPT Apps SDK and the Model Context Protocol (MCP). Explains how to add the app inside ChatGPT (open the app, add it, sign in to StockFit, ask a question), that no install or API key copying is needed, and that new accounts default to the free tier with no credit card. Covers what you can ask in plain language (revenue and margin trends, balance sheets, cash flow, EPS history, upcoming earnings dates, insider buying and selling via Form 4, institutional holders via 13F, ETF holdings and concentration, executive compensation, 10-K risk factors) and which StockFit endpoints back each answer (/api/financials/income-statement, /api/financials/balance-sheet, /api/earnings/snapshot, /api/insider-transactions, /api/fund/holdings, /api/company/research-summary). Explains that every value is parsed from SEC EDGAR filings with no third-party aggregators, the free tier scope versus paid plans (Starter $15/mo, Stock $39/mo, ETF $39/mo, Professional $69/mo), that upgrades happen on the StockFit website and the app picks up the new tier automatically, and that it is the same MCP server that works in Claude, Cursor, and VS Code on one account. Includes a 7-question FAQ.
- [Fundamental Stock Analysis Explained](https://developer.stockfit.io/blog/fundamental-stock-analysis): Field guide to fundamental stock analysis aimed at investors and developers who already know the term and want to understand how it is actually applied. Covers the five concrete use cases (valuation via DCF / comparables / sum-of-the-parts, stock screening, conviction and position sizing, value-trap and accounting-red-flag detection, credit and M&A diligence), the five-step practitioner workflow (understand the business, read the financial statements, quantify health and quality and growth, triangulate with ownership signals, decide and monitor), and the StockFit API endpoints that power each step: /api/company/details, /api/company/research-summary, /api/company/economic-model for business context; /api/financials/income-statement, /api/financials/balance-sheet, /api/financials/cash-flow-statement for the statements; /api/financials/key-metrics, /api/financials/scores, /api/financials/growth, /api/earnings/trends, /api/earnings/chart/quality for ratios and quality; /api/ownership/institutional-holders, /api/ownership/summary, /api/insider-transactions, /api/insider-transactions/summary for ownership signals; /api/earnings/snapshot, /api/earnings/eps-history, /api/earnings/dividend-history for monitoring. Discusses the four most common pitfalls (accruals vs cash flow, sector-mismatched ratios, fiscal-calendar drift, look-ahead bias from restated financials) and explains why point-in-time fundamentals data matters for backtesting. Includes a 5-question FAQ.
- [Insider Cluster Buys: What You Need to Know](https://developer.stockfit.io/blog/insider-cluster-buys-sp500): Event study testing the insider cluster buy signal across the S&P 500 from 2018 to 2024. Built from Form 4 transaction code P (open-market purchase) records pulled through /api/insider-transactions, with a discretionary-buyer filter that excludes DSPP and director-comp programmatic accumulation. 45 discretionary clusters, benchmarked against SPY via yfinance adjusted close at 30, 90, 180, and 365-day forward horizons. Findings: equal-weighted basket beats SPY at every horizon and peaks at +9.5% alpha around day 270; hit rate decays from 58% at 30 days to 42% at 365; mixed clusters (officers AND directors buying) deliver +7.8% mean while officers-only or directors-only clusters underperform; 2022 cohort dominates (CVNA +345%, GE +64%, FSLR +81%) and 2024 underperforms (ALB, DOW, CNC). Includes full TypeScript replication code, methodology footnote on survivorship bias, and named winners and losers with public Form 4 insider names.
- [Earnings Calendar API: Developer Guide](https://developer.stockfit.io/blog/earnings-preview-api): Developer guide to the StockFit Earnings Calendar API and the surrounding earnings toolkit. The earnings calendar API is three endpoints that all return the predicted earnings date, filing date, filing type, and (when available) the analyst consensus EPS estimate, the number of estimates, and expected report time (pre_market/after_hours): /api/earnings/upcoming (market-wide, next 1-28 days, paginated, sorted by date, each row carries a symbols[] array so multi-class issuers are one entry, plus the issuer name), /api/earnings/calendar (batch up to 50 tickers, returns a map keyed by symbol, delisted map to null), and /api/earnings/date (single ticker). To compute a beat or miss, compare epsEstimate against reported diluted EPS from /api/earnings/eps-history. Also covers the one-call executive summary (/api/earnings/snapshot), EPS history with basic and diluted weighted-average share counts and split-adjust (/api/earnings/eps-history), dividend history with payout ratio and coverage (/api/earnings/dividend-history, free tier), multi-year CAGRs and margin trajectory (/api/earnings/trends), line-item growth where annual = year-over-year and quarter = sequential (/api/financials/growth), series-oriented chart shapes with periods/series/rates axes (/api/earnings/chart/eps, /api/earnings/chart/quality), and how to watch the 8-K Item 2.02 print stream (/api/filings/latest?event=earnings, /api/filings/item with item=2.02). Live NVIDIA (NVDA) data with the next report expected August 26, 2026. Free-tier accessible. Also available as MCP tools for Claude, Cursor, and VS Code AI agents.
- [Understanding SEC Forms: A Guide to 10-K and More](https://developer.stockfit.io/blog/sec-forms-explained): Field guide to every SEC form the StockFit API ingests — what each filing is for, what data we extract from it, and which API endpoint serves it. Covers Form 10-K (annual report with income statement, balance sheet, cash flow statement), Form 10-Q (quarterly), Form 20-F (foreign private issuers), Form 8-K (material events with item codes), Form S-1 / EFFECT (IPO prospectus and effective date), Form 25-NSE (authoritative delisting), Forms 3/4/5 (insider transactions with transaction codes), Schedule 13D / 13G (5%+ beneficial ownership), Form 13F-HR (institutional holdings), Forms NPORT-P and N-CEN (fund holdings and operations), Forms N-1A / 485BPOS / 497 (fund prospectuses), Forms N-CSR / N-CSRS (fund shareholder reports including dividends), DEF 14A / DEFA14A (proxy statements with executive compensation), and ARS. Includes a 9-question FAQ.
- [StockFit vs Alpha Vantage: Fundamentals APIs](https://developer.stockfit.io/blog/stockfit-vs-alpha-vantage-fundamentals): Side-by-side comparison covering standardized statements, raw XBRL access, SEC filings access, ownership and insider data, executive compensation, ETF coverage, and pricing. Verifiable sources on both sides, no vendor spin.
- [ETF Analysis: Free-Tier to Forensic](https://developer.stockfit.io/blog/etf-deep-lens-stockfit-api): Walks SPY end-to-end through every fund endpoint. Starts at /api/fund/profile and /api/fund/holdings on the entry tiers, then moves into N-PORT and N-CEN power tools — flows, changes, overlap, fund health, fee analysis, service providers — for a forensic ETF view your brokerage page does not show. Live JSON, charts, composite playbook.
- [Stock Backtesting with the Holy Grail API](https://developer.stockfit.io/blog/holy-grail-api-stock-backtesting): Look-ahead bias is the silent killer of quant backtests. StockFit is the only commercial fundamentals API that ships a per-fact, per-amendment audit trail with `before` values out of the box on every tier. Walkthrough of the data model and the unwind algorithm.
- [Sector Metrics vs Financial Modeling Prep](https://developer.stockfit.io/blog/sector-industry-metrics-vs-fmp): Why one-shape-fits-all key-metrics endpoints silently fail for banks, insurers, and REITs. Three-tier metric model (general / sectorMetrics / industryMetrics), full sector and industry coverage tables, real responses for JPM (banking NIM), PGR (insurance combined ratio), PLD (REIT FFO/AFFO), and NVDA (semiconductor R&D intensity). Includes Node.js step-by-step screen tutorial and an honest feature comparison vs FMP.
- [StockFit vs sec-api.io: Financial Data APIs](https://developer.stockfit.io/blog/stockfit-vs-sec-api-io): Honest side-by-side comparison covering XBRL-to-JSON conversion strategies (curated vs raw concept names), insider trading data and deduplication, real cold-call response timings, and pricing. Verified facts on both sides with citations to public docs.
- [XBRL Financials to JSON: Guide for Backtesting](https://developer.stockfit.io/blog/xbrl-to-json-backtesting): Why SEC XBRL is hard to use directly, what "as-reported" really means, and how to assemble a clean point-in-time dataset for systematic strategies. Covers concept proliferation, dimensional facts, period classification, fiscal calendars, Q4 reconstruction, and a Node.js backtesting workflow against the StockFit API.

## Insights

AI-generated company and ETF deep dives built entirely from the StockFit API.

- [All insights](https://developer.stockfit.io/insights): Browse the full collection.
- [StockFit Earnings APIs: Key Insights and Features](https://developer.stockfit.io/insights/stockfit-earnings-api-deep-dive): A guided tour of eight StockFit earnings endpoints using Coca-Cola (KO) as the canvas: earnings calendar and date prediction, one-call snapshot, quarterly EPS history with YoY growth, dividend track record, multi-year trends and CAGRs, and earnings quality metrics. Maps each endpoint (/api/earnings/calendar, /api/earnings/date, /api/earnings/snapshot, /api/earnings/eps-history, /api/earnings/dividend-history, /api/earnings/trends, /api/earnings/chart/eps, /api/earnings/chart/quality) to the structured data it returns, all parsed from SEC filings.
- [StockFit Company APIs: Deep Dive Overview](https://developer.stockfit.io/insights/stockfit-company-api-deep-dive): A guided tour of the StockFit company endpoints using Apple (AAPL) as the canvas, turning SEC filings into structured business intelligence: company details and peers, the citation-verified economic model (business model, competitive advantages, flywheels, operating levers, failure modes), and research summary (/api/company/details, /api/company/peers, /api/company/economic-model, /api/company/research-summary).
- [NVIDIA (NVDA): The AI Infrastructure Colossus](https://developer.stockfit.io/insights/nvda-deep-dive): Financials, business model, and competitive moats of the world's most valuable semiconductor company.
- [Costco (COST): The Membership Machine](https://developer.stockfit.io/insights/cost-deep-dive): How a 2.9% net margin produces 37% returns on capital.
- [ARK Innovation ETF (ARKK): The Rise and the Fall](https://developer.stockfit.io/insights/arkk-deep-dive): Fund flows, holdings, sector shifts, and QQQ overlap analysis.

## Optional

- [Terms of Service](https://developer.stockfit.io/terms)
- [Privacy Policy](https://developer.stockfit.io/privacy)
- [Sitemap](https://developer.stockfit.io/sitemap.xml)