The Form F-3 Files Dataset is a complete archive of every Form F-3 and Form F-3/A registration statement submitted to EDGAR by foreign private issuers since October 1995, when EDGAR began accepting the form electronically. Form F-3 is the short-form Securities Act of 1933 registration statement available to seasoned foreign private issuers (FPIs, as defined in Rule 405) — the FPI counterpart to domestic Form S-3 — and Form F-3/A is its pre-effective or post-effective amendment. Each record in the dataset is a single EDGAR accession identified by its 18-digit accession number, packaged as a folder containing a metadata.json submission manifest and every textual document the filer transmitted (the primary registration statement plus exhibits such as underwriting agreements, securities instruments, legal opinions, consents, and the Exhibit 107 filing-fee table). The dataset covers both initial filings and all amendments under one shared schema, distributed as monthly ZIP containers with file types HTML, TXT, PDF, and JSON.
Programmatically retrieve the full list of dataset archive files, download URLs and dataset metadata.
Dataset Index JSON API
Download the entire dataset as a single archive file.
Download Entire Dataset:
Download a single container file (e.g. monthly archive) from the dataset.
Download Single Container:
Each record in the Form F-3 Files Dataset is a single EDGAR submission of either a Form F-3 (short-form registration statement under the Securities Act of 1933 by an eligible foreign private issuer) or a Form F-3/A (a pre-effective or post-effective amendment to a previously filed F-3). The unit of identity is the SEC accession number. Each record is a self-contained folder named with the 18-digit accession number with dashes stripped (for example 000110465925063822, corresponding to accession 0001104659-25-063822). Inside the folder are exactly one metadata.json describing the submission and every textual document the filer transmitted to EDGAR for that accession. One folder therefore corresponds one-to-one with one row in EDGAR's filing index, one entry in the issuer's registration history, and one 333- series Securities Act file number assignment (shared with sibling amendments). Records are packaged into monthly ZIP archives whose root directory is named for the period (for example 2025-06/), with one accession-number subfolder per submission filed in that month.
Form F-3 is the foreign-issuer counterpart to Form S-3. It is the short-form registration statement available to foreign private issuers that have a sufficient Exchange Act reporting history (generally at least twelve months of timely-filed reports) and that satisfy either the public-float test or one of the alternative transaction-based eligibility prongs (investment-grade non-convertible securities, secondary offerings by selling shareholders, rights offerings, dividend or interest reinvestment plans, etc.). Because eligibility presupposes a developed Exchange Act record, the form is "short" in that most issuer-level disclosure is satisfied by incorporation by reference to the filer's Form 20-F annual report and Form 6-K furnished interim reports rather than restated in the registration statement itself.
What the F-3 itself carries is concentrated on the offering: the Section 10 prospectus, the description of the securities being registered, the plan of distribution, offering-specific risk factors, the use of proceeds, the incorporation-by-reference legend, and the Part II registration-statement information (other expenses of issuance and distribution, indemnification of directors and officers, recent sales of unregistered securities, exhibits, undertakings, signatures). A Form F-3/A is filed to amend an F-3, whether to respond to staff comments, to file or refile an exhibit, to register additional securities, to update incorporated-by-reference information, or to reflect a change in the offering. Amendments may re-file the entire registration statement or only a narrow set of pages, exhibits, or cover material; both patterns appear in the dataset. The dataset spans October 1995 through the present, covering both initial filings and all amendments under one shared schema, and is distributed as monthly ZIP containers with file types HTML, TXT, PDF, and JSON.
A record has three concentric layers:
metadata.json — a single JSON object that mirrors the EDGAR submission header and indexes every document the issuer filed.metadata.json — the submission manifestEvery accession folder contains exactly one metadata.json, encoded as a single JSON object. It is the authoritative index for the record and the only reliable way to map files on disk to their EDGAR document role, because filename conventions cluster by filing agent rather than by document type.
Top-level fields:
formType — either "F-3" or "F-3/A".accessionNo — dashed accession number, e.g. "0001104659-25-063822". The parent folder name is the same number with dashes removed.filedAt — ISO 8601 acceptance timestamp with timezone offset, e.g. "2025-06-30T07:53:18-04:00".description — EDGAR human-readable form description, e.g. "Form F-3/A - Registration statement by foreign private issuers: [Amend]".linkToFilingDetails, linkToTxt, linkToHtml — absolute URLs to the primary document, the complete SGML submission text file, and the EDGAR -index.htm page on sec.gov/Archives/edgar/data/<cik>/<accessionNoNoDashes>/.linkToXbrl — typically an empty string for F-3 filings.id — a 32-character hex content hash for the submission.documentFormatFiles[] — the manifest of every file in the EDGAR submission. Each entry carries:
sequence — string "1", "2", ..., with the trailing complete-submission text-file entry using a single space " ".documentUrl — absolute URL to the file on sec.gov.description — free-text label, e.g. "FORM F-3/A", "EXHIBIT 5.1", "CONSENT", "FILING FEE TABLE", "GRAPHIC", "Complete submission text file". May be omitted on GRAPHIC rows.type — the EDGAR document-type code: "F-3", "F-3/A", "EX-1.x", "EX-4.x", "EX-5.1"/"EX-5.2", "EX-10.x", "EX-23.x", "EX-FILING FEES", "GRAPHIC", "XML", or " " for the wrapper submission.size — file size in bytes as a string.entities[] — normally a single-element array describing the filer. Each entity carries companyName suffixed with the role in parentheses (e.g. "Origin Agritech LTD (Filer)"), the numeric cik, the fileNo (a 333- series number for Securities Act registrations, e.g. "333-277955"), the EDGAR filmNo, an entity-level type mirroring the form type, the statute act ("33" for the Securities Act of 1933), the SIC industry code with label, the irsNo (often "000000000" for foreign issuers without a U.S. EIN), the fiscalYearEnd as MMDD, the stateOfIncorporation using EDGAR codes ("D6" Cayman Islands, "D8" British Virgin Islands, "X0" United Kingdom, "A6" Ontario, etc.), and a tickers array.seriesAndClassesContractsInformation — present but normally empty for F-3 filings (it is meaningful for fund and variable-product filings).dataFiles — present but normally empty, since F-3 submissions rarely carry XBRL data files.metadata.json serves two roles: it preserves header information that EDGAR otherwise exposes only inside the SGML wrapper of the submission, and it provides the authoritative crosswalk from on-disk filenames to EDGAR document types. Because filenames follow each filer-agent's house style (tm… for Toppan Merrill, ea… for EdgarAgents, d… for Donnelley, ticker-prefixed templates such as cney_f3.htm, generic templates formf-3.htm / formf-3a.htm, date-coded templates such as f3_062625.htm), the only robust way to locate the primary registration statement is to consult documentFormatFiles[] for the entry whose type is F-3 or F-3/A, conventionally at sequence "1".
Every non-JSON file in the folder is an EDGAR-formatted document: a small SGML header followed by the document body inside <TEXT>...</TEXT>. A representative envelope:
1
<DOCUMENT>
2
<TYPE>EX-23.1
3
<SEQUENCE>2
4
<FILENAME>cney_ex231.htm
5
<DESCRIPTION>CONSENT
6
<TEXT>
7
<html>... consent of independent registered public accounting firm ...</html>
8
</TEXT>
9
</DOCUMENT>
Header fields — <TYPE>, <SEQUENCE>, <FILENAME>, <DESCRIPTION> — mirror the corresponding entries in documentFormatFiles[]. The <TEXT> block contains the literal HTML the issuer filed; for F-3 records the body is consistently HTML (the file-types found in the dataset are TXT, JSON, HTML, and PDF, but modern F-3 records are overwhelmingly HTML plus the JSON manifest). The primary registration statement is by a wide margin the largest document in the folder, often several hundred kilobytes, while exhibits range from one-paragraph consents (1–3 KB) to long indenture, warrant, or sales-agreement forms (tens to hundreds of kilobytes).
For text extraction, only the contents of the <TEXT>...</TEXT> block should be parsed as HTML; the SGML header lines above and the closing tags below are not valid HTML and should be stripped first.
The primary document is the registration statement itself. It opens with the EDGAR cover boilerplate ("As filed with the Securities and Exchange Commission on <date>", "Registration No. 333-...", "UNITED STATES SECURITIES AND EXCHANGE COMMISSION", "FORM F-3 — REGISTRATION STATEMENT UNDER THE SECURITIES ACT OF 1933"), followed by issuer identification (exact name as specified in its charter, jurisdiction of incorporation, IRS Employer Identification Number where applicable, address of principal executive offices, and the U.S. agent for service of process required of foreign issuers). Pre-2022 cover pages also carry a Calculation of Registration Fee table; from 2022 onward that calculation is delivered as a separate Exhibit 107 (EX-FILING FEES).
The body is a Section 10 prospectus comprising:
The document closes with Part II information (other expenses, indemnification of directors and officers, recent sales of unregistered securities, exhibit index, undertakings) and the signature pages. Form F-3 General Instructions require the registration statement to be signed by the issuer, the principal executive officer, the principal financial officer, the principal accounting officer (or controller), a majority of the board of directors, and an authorized U.S. representative.
Forms of underwriting agreement, equity distribution agreement, at-the-market sales agreement, placement-agent agreement, or other distribution agreements. These exhibits embed commercial terms of the offering and frequently run tens to hundreds of kilobytes. They appear when the F-3 covers a takedown for which the agreement form is being filed contemporaneously rather than later under a Form 6-K.
Forms of warrant, indenture, supplemental indenture, deposit agreement (for ADSs), specimen share certificates, certificates of designation for preferred shares, and other instruments defining the rights of holders. Indenture-heavy shelf F-3s for debt issuers can carry six or more EX-4 sub-exhibits (for example EX-4.1 through EX-4.6), reflecting separate base, senior, and subordinated indentures plus their forms of note.
Opinions of counsel on the validity (and, for debt, the binding nature) of the securities being registered, required by Item 601(b)(5) of Regulation S-K. Foreign private issuers commonly file two: a local-jurisdiction opinion (Cayman Islands, BVI, Israeli, Bermudian, Marshall Islands, Irish, Luxembourg, etc.) addressing corporate authority and valid issuance under home-country law, and a U.S. counsel opinion addressing matters of New York law for debt or warrant securities. The opinions are addressed to the issuer, dated as of the filing date, and typically delivered as PDF-converted-to-HTML pages with a typed signature block.
Material contracts filed under Item 601(b)(10), present in a minority of F-3 records (for example, where a takedown is being made under a previously undisclosed standby commitment).
Short consents of independent registered public accounting firms and, when relevant, other named experts (petroleum reserve auditors, geologists, named tax counsel) authorizing inclusion or incorporation by reference of their reports. Auditor consents typically reference the audit firm's PCAOB ID (e.g. PCAOB ID #6907) and the specific 20-F annual report being incorporated. Filings that incorporate multiple audited entities, predecessor auditors, or expert reports can carry many consents (EX-23.x EX-23.1 through EX-23.9 and beyond). Each consent is short (1–3 KB).
The standardized HTML exhibit introduced by the SEC's October 2021 Filing Fee Disclosure rule, with mandatory compliance phased in through 2022. Before the rule, the fee calculation was a table on the cover page of the registration statement; after the rule, the calculation moved into a dedicated exhibit (Exhibit 107, EDGAR document type EX-FILING FEES) accompanied by a structured XML data file. The HTML table lists each security being registered (e.g. "Equity / Class A Ordinary Shares", "Debt / Senior Notes"), the fee-calculation rule invoked (Rule 457(c), Rule 457(o), Rule 457(r) for automatic-shelf takedowns, etc.), the amount registered, the maximum offering price per share, the maximum aggregate offering price, the fee rate (e.g. 0.0001531), and the registration fee due. Footnotes typically explain Rule 416(a) anti-dilution coverage and the basis for the per-share price estimate.
documentFormatFiles[] may also list GRAPHIC entries for image files (issuer logos, signature graphics, charts) and XML entries for the structured filing-fee data submitted alongside Exhibit 107. The trailing manifest entry whose sequence is a single space points to the EDGAR-generated <accession>.txt wrapper that concatenates the SGML envelopes of all documents.
On disk, each record contains metadata.json plus every textual document in the EDGAR submission: the primary F-3 or F-3/A registration statement and all numbered exhibits (EX-1.x, EX-4.x, EX-5.x, EX-10.x, EX-23.x, EX-FILING FEES, and the structured filing-fee XML where present). All textual documents are HTML wrapped in the EDGAR SGML envelope. Per-folder file count varies from a minimum of two (just metadata.json and a single primary document — common for narrow F-3/A amendments that re-file only the cover page) to well above ten for multi-exhibit shelf registrations with many consents and instruments.
GRAPHIC entries, typically *.jpg) are listed in documentFormatFiles[] but are excluded from the on-disk record by design.<accession>.txt wrapper that EDGAR distributes (a concatenation of every document's SGML envelope) is indexed in documentFormatFiles[] and reachable through linkToTxt, but not duplicated as a file inside the folder, since the individual documents already carry the same SGML headers.Form F-3 has existed since 1982 as the foreign-issuer analog of Form S-3. EDGAR began accepting it electronically in October 1995, which is why the dataset's earliest records are dated 1995-10. Several rounds of rulemaking have reshaped the content over the years.
The most consequential change is the 2005 Securities Offering Reform. It introduced the Well-Known Seasoned Issuer (WKSI) concept and the automatic shelf registration regime, allowing eligible WKSIs (including qualifying foreign private issuers) to register an unspecified amount of securities on Form F-3 with immediate effectiveness on filing under Rule 462(e), defer payment of registration fees to the time of takedown under Rules 456(b) / 457(r), add new classes of securities by post-effective amendment, and rely on a broader concept of free writing prospectus under Rules 163, 163A, 164, and 433. Pre-2005 F-3 prospectuses tend to register a specific dollar amount of securities with full fee payment on filing; post-2005 automatic-shelf F-3s often register classes generically, and EDGAR descriptions for such filings sometimes read "automatic shelf registration statement".
The 2008 amendments tightened eligibility so that issuers under investigation or with recent waiver-related disqualifications had narrower access. The 2018 Regulation S-K modernization removed some duplicative cover-page items and broadened incorporation by reference for smaller reporting companies, with parallel updates relevant to F-3. In 2020 the SEC eliminated the standalone ratio-of-earnings-to-fixed-charges exhibit requirement, so later filings drop that exhibit.
The most visible recent change is the SEC's October 2021 Filing Fee Disclosure rule, introducing the standalone Calculation of Filing Fee Tables exhibit (Exhibit 107, document type EX-FILING FEES) plus its companion structured XML data file. The rule moved fee calculations off the cover page into a dedicated exhibit. F-3 records filed before the rule's effectiveness generally do not include EX-FILING FEES; records filed after include it as a near-universal feature.
F-3 records reflect EDGAR's own format history. The earliest records, from 1995 through roughly the late 1990s, were filed as plain ASCII text inside the EDGAR SGML envelope: the <TEXT> block contained line-wrapped uppercase-heavy prose with limited tables. Through the late 1990s and 2000s, issuers transitioned to HTML inside the same SGML envelope, and by the mid-2000s HTML became the dominant body format for both the primary registration statement and exhibits. Modern F-3 and F-3/A records are uniformly HTML inside the SGML envelope, which is why nearly every non-JSON document file in this dataset carries an .htm or .html extension and an HTML body. PDF and plain-text bodies remain technically permitted for some attachments but are rare in practice for F-3 filings.
XBRL has not been a structural part of F-3 itself — the form does not carry primary financial statements, since those flow through the incorporated 20-F — so linkToXbrl is typically empty and dataFiles is typically an empty array. The one structured-data element that does appear in modern F-3 records is the XML companion to Exhibit 107, listed in documentFormatFiles[] with type of XML for filings made under the post-2022 Filing Fee Disclosure rule.
Filename conventions are not canonical. Filenames cluster by filing agent (Toppan Merrill tm…, EdgarAgents ea…, Donnelley d…, ticker-prefixed templates such as cney_f3.htm, generic templates formf-3.htm / formf-3a.htm, date-coded templates such as f3_062625.htm) and bear no canonical relationship to document role. Consumers should resolve documents through documentFormatFiles[], matching on type and sequence. The primary registration statement is consistently at sequence "1" with type of F-3 or F-3/A.
Amendments may be narrow. F-3/A submissions do not necessarily re-file the entire registration statement. Many amendments file only a revised cover page, a single updated exhibit, or a new consent, and the folder may contain as few as two files (metadata.json plus one primary HTML document). Reconstructing a complete amended registration statement requires reading the F-3/A together with the original F-3 it amends, which can be located via the shared 333- file number in entities[].fileNo.
Incorporation by reference is fundamental. The substantive issuer-level disclosure that would appear in a Form F-1 (the long-form registration statement) is largely absent from F-3, which leans on the most recent Form 20-F and subsequent Form 6-K filings. A reader looking for full historical financial statements will not find them inside an F-3 record; they live in the corresponding 20-F record.
Stitching keys. The 333- file number in entities[].fileNo is the natural key for grouping an F-3 family (the original F-3 and all of its F-3/A amendments) across the dataset. The cik is stable across all of an issuer's filings; fileNo is stable within a single registration-statement family. The accessionNo is unique to each individual submission.
SGML vs HTML boundary. The SGML wrapper around each document carries header metadata (<TYPE>, <SEQUENCE>, <FILENAME>, <DESCRIPTION>) that duplicates information in documentFormatFiles[]. For text extraction, parse only the contents of <TEXT>...</TEXT> as HTML and discard the surrounding SGML lines.
Automatic shelf takedowns are not in this dataset. For F-3 records made under Rule 462(e), the registration statement becomes effective immediately on filing. Subsequent takedowns under such a shelf are filed as Form 424B prospectus supplements (and any free writing prospectuses as Form FWP), which are separate EDGAR submissions and outside the scope of this dataset.
Each record is a Form F-3 registration statement (or Form F-3/A amendment) submitted to EDGAR by an eligible foreign private issuer (FPI) to register a securities offering under the Securities Act of 1933. The legal filer is the registrant itself: an issuer organized outside the United States that meets the FPI definition in Securities Act Rule 405 and the seasoning and transaction tests in Form F-3's General Instruction I. F-3/A records are amendments filed by the same registrant, typically in response to staff comments, to update financial or transactional disclosure, or as post-effective amendments to a shelf.
Form F-3 is the FPI counterpart to Form S-3. Domestic issuers use S-3; qualifying FPIs use F-3. FPIs that do not qualify fall back to Form F-1 (long-form), Form F-4 (business combinations), or Form F-10 (Canadian MJDS); none of those forms are in this dataset.
Two cumulative conditions bound the population.
Foreign private issuer status (Rule 405). A foreign issuer (other than a foreign government) qualifies as an FPI unless more than 50% of its outstanding voting securities are held of record by U.S. residents AND any of the following also applies: (i) the majority of executive officers or directors are U.S. citizens or residents, (ii) more than 50% of the issuer's assets are located in the United States, or (iii) the business is administered principally in the United States. FPI status is tested annually as of the last business day of the second fiscal quarter. Issuers failing the test become domestic registrants and must use S-3 instead of F-3.
Form F-3 eligibility (General Instruction I). Both prongs must be satisfied:
Filers include large multinational operating companies whose ordinary shares trade as ADRs, foreign banks and financial institutions, foreign sovereign-controlled commercial entities organized as private issuers, and smaller-cap FPIs using the baby shelf. Investment companies registered under the Investment Company Act do not use F-3; they file on N-1A, N-2, and related forms. Canadian issuers eligible for MJDS typically use Form F-10.
Form F-3 is event-driven, not periodic. A filing arises whenever an eligible FPI elects to register an offering on the short form. Common triggers:
F-3/A amendments are triggered by (i) staff comments during review of a non-automatic filing; (ii) pre-effective updates to financial statements, risk factors, or transaction terms; (iii) pre-effective inclusion of pricing or deal-specific information that cannot be added by a Rule 424 supplement; and (iv) post-effective amendments that add new securities, restate the registration statement, or update incorporated documents in ways a prospectus supplement or Section 10(a)(3) update cannot accomplish.
Form F-3 has no periodic deadline; timing follows the issuer's transaction calendar subject to effectiveness rules:
Form F-3 was adopted in 1982 as part of the integrated disclosure system (Securities Act Release No. 6383), which established the parallel S-1/S-2/S-3 (domestic) and F-1/F-2/F-3 (FPI) hierarchies. Before 1982, FPIs registered on predecessor forms. EDGAR mandated electronic filing for Securities Act registrations in phases between 1993 and 1996; the dataset's earliest sample date of October 1995 reflects this phase-in. The 2005 Securities Offering Reform introduced the WKSI/automatic-shelf branch, which from December 2005 onward accounts for a growing share of large-FPI Form F-3 filings.
Form F-3 sits among Securities Act registration forms and the Exchange Act reports it incorporates. The most useful comparisons are: (1) sibling registration forms (S-3, F-1, F-4, F-10, S-1), (2) the periodic FPI reports F-3 references (20-F, 6-K), (3) the takedowns it enables (424B), and (4) the F-3ASR variant.
S-3 is the structural twin of F-3: same short-form concept, parallel eligibility (timely Exchange Act reporting plus public-float or transaction-based tests), same building blocks (base prospectus, plan of distribution, risk factors, incorporation by reference). The split is the filer population. S-3 is filed by U.S. domestic registrants and incorporates 10-K, Form 10-Q, and Form 8-K. F-3 is filed only by Rule 405 foreign private issuers and incorporates 20-F and 6-K. F-3 records also carry FPI-specific items absent from S-3: home-country signing-officer blocks, IFRS or US-GAAP-reconciled financials by reference, and FPI governance disclosures. Combine the two only if the project requires a unified seasoned-issuer shelf universe; otherwise they are not interchangeable.
F-1 is what an FPI files when it cannot use F-3 (typically pre-IPO, sub-twelve-month reporting history, or below float thresholds). F-1 is self-contained: it embeds the full 20-F-equivalent item set, audited financials, and complete business description inside the registration statement. F-3 omits all of that and points to a previously filed 20-F plus intervening 6-Ks. F-1 records are large and disclosure-heavy; F-3 records are short transactional shells whose substance lives upstream.
F-4 registers securities issued in mergers, exchange offers, and share-for-share acquisitions involving FPIs. Overlap with F-3 is limited to the Securities Act registration mechanic and the option to incorporate by reference. F-4 carries items F-3 does not: target-company financials and disclosure, deal background, fairness opinions, voting and tax disclosure, and the merger or exchange agreement as an exhibit. F-3 carries shelf-program items F-4 does not: a multi-security base prospectus, generic plan of distribution, and no counterparty transaction. Use F-4 for M&A registration; use F-3 for capital-raising shelves.
F-10 is a short-form registration available only to qualifying Canadian issuers under the multijurisdictional disclosure system. It performs F-3's economic role but uses Canadian-prepared documents, Canadian continuous-disclosure incorporation (AIF, MD&A, Canadian interims), and Canadian GAAP/IFRS as accepted under MJDS rules rather than 20-F/6-K incorporation. F-10 is not in this dataset; cross-border short-form coverage of Canadian issuers requires F-10 separately. Many Canadian issuers prefer F-10 for the streamlined Canadian MJDS review.
Form S-1 is the domestic counterpart to F-1: full long-form Securities Act registration for U.S. domestic issuers without short-form eligibility, and the standard domestic IPO vehicle. The S-1 versus F-3 contrast is two-dimensional: domestic versus FPI, and self-contained long form versus reference-driven short form. F-3 is essentially never an IPO vehicle; it supports follow-on and shelf offerings by seasoned FPIs.
20-F is the Exchange Act annual report for FPIs and the principal document an F-3 incorporates by reference. 20-F is periodic, filed under the Exchange Act, and contains the full annual package: business description, risk factors, MD&A, audited IFRS or US-GAAP-reconciled financials, governance, and Item 16 disclosures. F-3 is event-driven, filed under the Securities Act, and contains transactional disclosure (offering plan, security terms, use of proceeds) without restating the annual content. For fundamentals use 20-F; for offering activity use F-3.
6-K is the FPI vehicle for furnishing material information already made public abroad: interim financials, press releases, and material event disclosures. F-3 incorporates 6-Ks by reference to update the prospectus between 20-Fs. 6-Ks are short, frequent, and furnished (not filed); F-3 is a single registration statement that references them. For interim and event-level disclosure use 6-K; for registered offering structure use F-3.
424B filings (424B1 through 424B8) are the prospectus supplements that effect individual takedowns off an effective F-3 shelf. They sit downstream: F-3 establishes the shelf and registered securities; each 424B specifies the takedown's size, pricing, underwriters, use of proceeds, and final terms. The F-3 dataset captures the shelf framework and registered capacity; a 424B dataset captures the actual offering events. They link by accession lineage and should be combined for end-to-end offering analytics, not treated as substitutes.
F-3ASR is not a separate form but a filing designation used when a well-known seasoned issuer (WKSI) submits an automatic shelf on Form F-3. F-3ASR filings become effective on filing without staff review and may register an unspecified amount of securities. The document is still an F-3 and is included in this dataset; the ASR distinction is an effectiveness and pay-as-you-go fee mechanic, not a separate disclosure regime. WKSI versus non-WKSI analysis must use form-type tags and effectiveness metadata within this dataset rather than a separate F-3ASR source.
The Form F-3 Files Dataset is narrowly scoped to short-form Securities Act registration by seasoned FPIs. It is not periodic disclosure (use 20-F or 6-K), not transactional M&A registration (use F-4), not long-form registration (use F-1 or S-1), not Canadian MJDS (use F-10), not domestic short-form (use S-3), and not takedown-level pricing data (use 424B). Its distinctive content — base prospectus, security description, plan of distribution, risk factors, and exhibits anchored to incorporation by reference of the FPI's 20-F and 6-K stream — exists only here. F-3 is most analytically powerful when joined upstream to 20-F and 6-K and downstream to 424B; together those four sources reconstruct an FPI's full registered-offering pipeline.
The dataset serves teams working on cross-border shelf registration for foreign private issuers: the lawyers drafting and reviewing the filing, the bankers structuring takedowns, the analysts pricing dilution risk, the compliance teams owning the record, and the data and research groups building tooling around it.
Issuer-side and underwriter-side counsel use the corpus as a precedent library. Associates filter metadata.json by entities (issuer jurisdiction, SIC) and formType (F-3 vs. F-3/A) to assemble peer sets, then open the base prospectus for cover page, description of securities, plan of distribution (ATM, forward sale, underwritten shelf), selling-securityholder tables, and use-of-proceeds language. They pull EX-5 legal opinions, EX-8 tax opinions, and EX-23 auditor consents to compare scope and qualifications, and read EX-FILING FEES (Exhibit 107) to confirm fee calculation on unallocated shelves and Rule 457 carryforwards. F-3/A diffs against the original F-3 surface staff comment patterns. Output: precedent memo and drafting checklist.
Coverage and syndicate bankers structure takedowns and benchmark mechanics. They read the plan of distribution for ATM, equity distribution agreement, forward-sale, and MTN structures, and the cover page plus fee table for shelf capacity, WKSI/automatic-shelf treatment, and unsold carryforward. filedAt and entities drive a competitor-shelf monitor that feeds pitch books on peer timing, size, and structure.
Sell-side and buy-side analysts treat new F-3s as forward indicators of issuance. Equity analysts read the cover page and prospectus summary for maximum aggregate offering amount, then watch downstream supplements to model ATM utilization, share-count impact, and proceeds use. Credit analysts focus on the description of debt securities, indenture exhibits, and guarantor structure. Both reconcile against the incorporated 20-F and any 6-Ks listed in the incorporation-by-reference section. Output: dilution model, ATM tracker, updated credit view.
In-house securities counsel benchmark their own filings against home-jurisdiction and industry peers. They compare incorporation-by-reference language, risk factor evolution across successive F-3s, forward-looking-statements safe-harbor wording, enforceability-of-civil-liabilities sections, and currency-translation disclosures. The workflow asks whether risk factors are current, the 6-K incorporation list is complete, and EX-5, EX-23, and EX-FILING FEES match current market practice. Output: internal disclosure review and comment log.
These desks treat new F-3s as leading indicators of imminent capital raises. They poll filedAt and entities from metadata.json to detect new shelves, then read the plan of distribution and use-of-proceeds language to infer opportunistic ATM, marketed follow-on, convertible, or debt refinancing. Distressed-credit teams focus on F-3s from leveraged FPIs, where the mix of registered securities shapes restructuring expectations. Output: issuance alerts and trade theses.
Researchers studying cross-listed firms, the MJDS for Canadian issuers, WKSI and automatic shelves, and the 2005 Securities Offering Reform use the 1995-to-present corpus as a longitudinal panel. Typical questions: effect of WKSI eligibility on shelf usage, MJDS vs. non-MJDS shelf size and takedown frequency, IFRS vs. US GAAP vs. home GAAP and shelf structure. Inputs are the cover page, fee table, exhibit list, and incorporation-by-reference section. Output: working papers and regulatory submissions.
Engineering teams build issuance pipelines that link base shelves to subsequent prospectus supplements via accessionNo and entities, parse EX-FILING FEES for registered amounts and carryforwards, and emit dilution alerts when the cover page or plan of distribution indicates new ATM or follow-on capacity. documentFormatFiles lets pipelines locate the prospectus, EX-5, EX-23, EX-FILING FEES, and indenture exhibits without parsing full submissions. Quants use the historical corpus to calibrate base rates on time-from-shelf-to-takedown, capacity utilization, and post-shelf return patterns. Output: issuance feed, dilution model, factor signals.
Analysts at regulators, SROs, and policy groups track how F-3 use has shifted since the 2005 Offering Reform and across crisis periods, and how staff review patterns differ across industries and jurisdictions (visible in F-3/A amendments). They use the cover page for filer status (WKSI, smaller reporting company, accelerated filer), the fee table for offering-size distributions, and the exhibit list for mandated-disclosure changes. Output: rulemaking memos and supervisory analytics.
Teams building RAG and drafting assistants for capital-markets workflows use the corpus to index plan-of-distribution language, FPI risk factors, EX-5 opinion templates, and shelf-mechanics Q&A. Three decades of F-3 and F-3/A filings give retrieval systems coverage of standard structures, regulatory-driven language changes, and issuer-specific variation. Output: RAG retrieval index and fine-tuning corpus.
The use cases below describe concrete workflows built directly on F-3 and F-3/A records, the metadata.json manifest, and the SGML-wrapped exhibits.
Precedent search across past F-3s by jurisdiction and SIC for drafting. Issuer-side and underwriter-side associates filter metadata.json on entities[].stateOfIncorporation (e.g. D6 Cayman, D8 BVI, X0 UK), entities[].sicCode, and formType to assemble a peer set, then extract the description-of-securities, plan-of-distribution, and use-of-proceeds blocks from the sequence-1 primary document. Output: a precedent memo and clause bank tagged by jurisdiction, SIC, and offering type that anchors the first draft of a new shelf.
Shelf-takedown tracker stitching the 333- family. Engineering teams group every F-3 and F-3/A by entities[].fileNo (the 333-... Securities Act file number) and join out-of-dataset 424B prospectus supplements on the same file number. The result is a per-shelf timeline showing original effectiveness, each amendment, and downstream takedowns. Deliverable: an issuance feed for ECM bankers and event-driven desks that flags newly effective shelves and measures time-from-shelf-to-takedown by issuer and structure.
WKSI vs non-WKSI population segmentation. Analysts split the corpus by EDGAR description (filings labeled "Automatic shelf registration statement") and Exhibit 107 footnote markers for Rule 457(r) pay-as-you-go to build WKSI and non-WKSI cohorts. Cross-tabulating against entities[].stateOfIncorporation, SIC, and filedAt produces longitudinal panels for studying how the 2005 Securities Offering Reform shifted FPI shelf usage. Deliverable: a labeled panel dataset for working papers and rulemaking memos.
FPI dilution monitor from cover page and Exhibit 107. Equity research teams pull filedAt and entities[].cik from metadata.json, then parse the cover page and EX-FILING FEES exhibit for maximum aggregate offering amount, securities classes, and Rule 457 carryforwards. The signal is fed into share-count and ATM-utilization models for cross-listed FPIs. Deliverable: a dilution dashboard with per-ticker registered capacity and rolling utilization estimates.
Exhibit 107 fee-table extraction for offering-size analytics. Quants and policy researchers parse the EX-FILING FEES HTML and its companion XML (post-2022 records) to extract security type, fee-calculation rule (457(c), 457(o), 457(r)), amount registered, max aggregate offering price, fee rate, and footnoted carryforwards. Aggregating across the corpus yields offering-size distributions by SIC, jurisdiction, and year. Deliverable: a normalized fee-table table powering market-sizing reports and supervisory analytics.
Risk-factor evolution corpus per issuer. In-house FPI counsel and academic researchers stitch successive F-3s for the same cik ordered by filedAt, isolate the offering-specific risk factors and forward-looking-statements safe-harbor language in the primary document, and diff them across filings. Deliverable: a per-issuer risk-factor changelog and a cross-issuer corpus for benchmarking enforceability-of-civil-liabilities, currency-translation, and home-country disclosures.
RAG retrieval index over base prospectus, EX-5 opinions, and EX-23 consents. LLM teams parse only the <TEXT>...</TEXT> block of each SGML envelope, chunk by registration-statement section (forepart, risk factors, plan of distribution, description of securities, Part II) and by exhibit type, and tag each chunk with cik, fileNo, formType, jurisdiction, and filedAt. Joining incorporated-by-reference 20-F filings completes the context. Deliverable: a retrieval index that powers drafting assistants and shelf-mechanics Q&A across three decades of FPI filings.
Cross-listed FPI universe and counsel network from entities[] and EX-5. Researchers and BD teams build a master list of seasoned FPIs by deduplicating entities[].cik across all F-3 records and joining stateOfIncorporation, sicCode, tickers, and fiscalYearEnd. Parsing the signature blocks of EX-5.1 and EX-5.2 legal opinions yields the local-jurisdiction and U.S. counsel firms used by each issuer. Deliverable: a cross-listed FPI universe with mapped counsel relationships for league tables and outreach lists.
F-3/A amendment-pattern analytics for staff comment inference. Compliance and regulatory analysts isolate formType = "F-3/A" records, group them with the originating F-3 via entities[].fileNo, and characterize each amendment by the document set re-filed (cover only, single exhibit, full re-file) using documentFormatFiles[].type. Aggregating by SIC and jurisdiction reveals where staff review concentrates. Deliverable: a heatmap of amendment intensity used to calibrate review-time expectations and to brief drafting teams on likely comment areas.
Indenture and securities-instrument library from EX-4 exhibits. DCM bankers and credit analysts filter records whose documentFormatFiles[] contains EX-4.1 through EX-4.x entries indicating base, senior, or subordinated indentures and forms of note, then extract the indenture and supplemental-indenture text. Deliverable: an indenture clause library indexed by issuer jurisdiction and security type, supporting covenant benchmarking and guarantor-structure comparisons across FPI debt shelves.
The Form F-3 Files Dataset is distributed through three endpoints: a public JSON index for metadata discovery, a single archive download for the full dataset, and per-container downloads for incremental retrieval. Authenticated endpoints accept the API key either as a ?token=YOUR_API_KEY query parameter or as a standard Authorization header.
Dataset Index JSON API: https://api.sec-api.io/datasets/form-f3-files.json
Returns dataset-level metadata (name, description, updatedAt, earliestSampleDate, totalRecords, totalSize, formTypes, containerFormat, fileTypes), the full-archive download URL, and the list of container files with per-container key, size, records, updatedAt, and downloadUrl. Use this endpoint to monitor which containers were rewritten in the latest refresh and download only those, rather than re-fetching the full archive. This endpoint does not require an API key.
1
curl https://api.sec-api.io/datasets/form-f3-files.json
Example response:
1
{
2
"datasetId": "1f13365b-9ae0-6933-8333-8e5237c68b01",
3
"datasetDownloadUrl": "https://api.sec-api.io/datasets/form-f3-files.zip",
4
"name": "Form F-3 Files Dataset",
5
"updatedAt": "2026-05-07T02:50:23.422Z",
6
"earliestSampleDate": "1995-10-01",
7
"totalRecords": 20608,
8
"totalSize": 448099597,
9
"formTypes": ["F-3", "F-3/A"],
10
"containerFormat": "ZIP",
11
"fileTypes": ["TXT", "JSON", "HTML", "PDF"],
12
"containers": [
13
{
14
"downloadUrl": "https://api.sec-api.io/datasets/form-f3-files/2026/2026-05.zip",
15
"key": "2026/2026-05.zip",
16
"size": 4218947,
17
"records": 38,
18
"updatedAt": "2026-05-07T02:50:23.422Z"
19
}
20
]
21
}
Download Entire Dataset: https://api.sec-api.io/datasets/form-f3-files.zip?token=YOUR_API_KEY
Downloads the complete dataset as a single ZIP archive covering all F-3 and F-3/A filings since October 1995. This endpoint requires an API key.
Download Single Container: https://api.sec-api.io/datasets/form-f3-files/2026/2026-05.zip?token=YOUR_API_KEY
Downloads one container (typically a monthly archive) using the key value from the index response. Each container ZIP unpacks into one folder per accession number, where each folder holds the original EDGAR submission documents — the primary HTML document, exhibits (HTML, TXT, PDF), and a metadata.json describing the filing. Image files (.jpg, etc.) are excluded from the containers but are referenced in metadata.json for retrieval from EDGAR if needed. This endpoint requires an API key.
The dataset covers Form F-3 and Form F-3/A filings submitted to EDGAR. Form F-3 is the short-form Securities Act of 1933 registration statement available to eligible foreign private issuers, and Form F-3/A is its pre-effective or post-effective amendment. Other FPI registration forms — F-1, F-4, F-10, F-3D, F-3DPOS — are not included.
S-3 and F-3 are structural twins with the same short-form concept, parallel eligibility (timely Exchange Act reporting plus public-float or transaction-based tests), and the same building blocks (base prospectus, plan of distribution, risk factors, incorporation by reference). The difference is the filer population: S-3 is filed by U.S. domestic registrants and incorporates Forms 10-K, 10-Q, and 8-K, while F-3 is filed only by Rule 405 foreign private issuers and incorporates Forms 20-F and 6-K. An issuer that loses FPI status must transition from F-3 to S-3.
F-1 is the long-form FPI registration statement and is self-contained: it embeds the full 20-F-equivalent item set, audited financials, and complete business description inside the registration statement. F-3 is the short form available only to seasoned FPIs that have at least twelve months of timely Exchange Act reporting and meet the public-float or transaction tests; it omits issuer-level disclosure and points instead to a previously filed 20-F plus intervening 6-Ks. First-time U.S. registrants typically begin on F-1 and graduate to F-3 once they qualify.
F-3/A denotes an amendment to a previously filed Form F-3. Amendments are filed to respond to staff comments, to update financial or transactional disclosure pre-effectiveness, to add securities or restate disclosure post-effectiveness, or to file or refile an exhibit. Amendments may re-file the entire registration statement or only a narrow set of pages or exhibits; both patterns appear in the dataset, and the original F-3 and all of its F-3/A amendments share a common 333- series file number in entities[].fileNo.
The substantive issuer-level disclosure that would otherwise appear in a long-form registration statement — historical financial statements, MD&A, business description, comprehensive risk factors, governance — is incorporated by reference rather than restated. F-3 pulls in the issuer's most recent Form 20-F annual report, all Form 6-K furnishings since the 20-F, and any other Exchange Act reports the issuer lists. Those incorporated 20-F and 6-K filings are not part of this dataset; only the F-3 and F-3/A submissions themselves are included.
A Well-Known Seasoned Issuer (WKSI) is an issuer that meets a higher-tier seasoning standard under Rule 405 — generally at least US$700 million in non-affiliate common equity float, or US$1 billion of registered non-convertible debt issued in the preceding three years. WKSIs may file an automatic shelf on Form F-3 that becomes effective immediately upon filing under Rule 462(e), without staff review, and may register an unspecified amount of securities with pay-as-you-go fees under Rules 456(b) / 457(r). These automatic shelves are sometimes labeled "F-3ASR" but remain Form F-3 filings and are part of this dataset; their EDGAR description often reads "Automatic shelf registration statement".
Subsequent takedowns under an effective F-3 shelf are filed as Form 424B prospectus supplements (424B1 through 424B8) and any free writing prospectuses are filed as Form FWP. These are separate EDGAR submissions outside the F-3/F-3/A scope and are not part of this dataset. To reconstruct an end-to-end offering pipeline, join the F-3 dataset to a 424B dataset on the shared 333- file number in entities[].fileNo.
metadata.json and what file types are in each container?Every accession folder contains exactly one metadata.json JSON object with top-level fields including formType, accessionNo, filedAt, description, linkToFilingDetails, linkToTxt, linkToHtml, linkToXbrl, id, documentFormatFiles[] (the manifest of every file with sequence, documentUrl, description, type, and size), entities[] (issuer companyName, cik, fileNo, filmNo, act, sicCode, irsNo, fiscalYearEnd, stateOfIncorporation, and tickers), seriesAndClassesContractsInformation, and dataFiles. Containers are monthly ZIP archives, and the file types found inside are HTML, TXT, PDF, and JSON. Image files (GRAPHIC entries such as .jpg) are listed in documentFormatFiles[] but excluded from the on-disk record by design.