The Form S-3 Files Dataset is a file-level archive of every Form S-3 and Form S-3/A registration statement accepted by EDGAR, packaged accession by accession with the original submission documents intact. Each record is a single registration statement filed under the Securities Act of 1933 by a U.S. domestic seasoned reporting issuer — a registrant that meets the General Instruction I.A reporting and currency tests and at least one General Instruction I.B transaction test (commonly the USD 75 million public-float threshold for primary offerings, the listed-class condition for secondary resales, or the WKSI thresholds for automatic shelves). A record bundles the per-submission metadata.json together with the primary S-3 (or S-3/A) document and every non-image exhibit transmitted with it — legal opinions, auditor consents, indentures, underwriting agreements, fee tables, and supporting exhibits. The dataset begins with EDGAR submissions from 1994-01-01 and is refreshed on a monthly cadence, distributed as monthly ZIP containers organized by year, with constituent files preserved in their original HTML, TXT, JSON, and PDF formats.
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:
The dataset contains the as-filed bundle of every Form S-3 and Form S-3/A registration statement on EDGAR. Form S-3 is the short-form registration statement under the Securities Act of 1933 used by seasoned reporting companies to register securities offerings. Its defining feature is heavy reliance on incorporation by reference: rather than restating the registrant's full business, financial, and management disclosure inside the registration statement, Form S-3 imports those disclosures from the registrant's existing Exchange Act reports (10-K, 10-Q, 8-K, and proxy statements). The filing tends to be relatively thin in narrative pages compared with Form S-1 but carries the same legal force as a full registration statement and is supported by the prospectus delivery framework of the 1933 Act.
Form S-3 is the standard vehicle for follow-on equity offerings, debt shelf registrations, medium-term note programs, at-the-market (ATM) equity programs, dividend reinvestment plans, resale registrations on behalf of selling securityholders, and — for well-known seasoned issuers (WKSIs) — automatic shelf registrations (ASRs) that become effective immediately upon filing. Each of these use cases produces an S-3 with the same broad anatomy but different emphases in the prospectus body, the plan of distribution, and the exhibit list.
The dataset is distributed as monthly ZIP containers organized by year (for example, 2025/2025-06.zip). Each container holds one top-level directory per filing accession. The file types found in the dataset are HTML/HTM (the dominant format for the primary document and exhibits), JSON (the per-record metadata.json), TXT (legacy ASCII bodies inside the SGML envelope and very early-EDGAR submissions), and PDF (used for certain exhibit classes and for permitted unofficial copies). Coverage starts at 1994-01-01 and continues forward with each refresh.
One record in the Form S-3 Files Dataset is a single Form S-3 or Form S-3/A registration statement as accepted by EDGAR, captured at the granularity of one accession number. Physically, a record is a folder named with the 18-digit EDGAR accession number with dashes removed (for example, accession 0001193125-25-140744 becomes folder 000119312525140744). The folder holds a metadata.json describing the submission together with one preserved file per non-image component of the original EDGAR filing — the primary S-3 (or S-3/A) registration document plus every exhibit that was transmitted with it. The dataset therefore preserves the as-filed bundle for each registration statement, accession by accession, with image binaries omitted.
A record corresponds either to an initial registration (S-3) or to an amendment (S-3/A). Both share the same internal anatomy; amendments commonly carry a narrower exhibit set (often only refreshed auditor consents and a re-issued legal opinion) and an explanatory note on the cover identifying the prior file number and the scope of the amendment.
Each accession folder inside a monthly container contains:
metadata.json describing the EDGAR submission, its registrant entities, and an inventory of every component document; and<prefix>s3.htm, <prefix>s3a.htm, <prefix>ex5_1.htm, <prefix>ex23_1.htm, <prefix>exfilingfees.htm, and filer-specific variants such as forms-3.htm or ex107.htm.Each shipped HTML or text document retains the EDGAR SGML document envelope. The file begins with header lines of the form
1
<DOCUMENT>
2
<TYPE>S-3
3
<SEQUENCE>1
4
<FILENAME>d50311ds3.htm
5
<DESCRIPTION>S-3
6
<TEXT>
7
<HTML> ... </HTML>
8
</TEXT>
9
</DOCUMENT>
The <TYPE>, <SEQUENCE>, <FILENAME>, and <DESCRIPTION> lines mirror the corresponding entry in metadata.json's documentFormatFiles[], and the <TEXT> block contains the actual HTML body that a browser would render. Exhibits use the same envelope with the appropriate exhibit <TYPE> such as EX-5.1, EX-23.1, or EX-FILING FEES.
metadata.json shapeThe metadata file is a single JSON object capturing what EDGAR knew about the submission at acceptance time. The top-level fields are:
formType — "S-3" for new registration statements, "S-3/A" for amendments.accessionNo — EDGAR accession in dashed canonical form (e.g. "0001193125-25-140744").description — EDGAR's human description, typically "Form S-3 - Registration statement under Securities Act of 1933" (suffixed with [Amend] for S-3/A).filedAt — ISO 8601 acceptance timestamp with EDGAR's Eastern offset.linkToFilingDetails, linkToTxt, linkToHtml, linkToXbrl — URLs to the primary HTML document, the full SGML .txt submission, the EDGAR -index.htm page, and any XBRL instance respectively. For S-3 the XBRL link is typically empty because S-3 itself does not carry financial-statement XBRL.documentFormatFiles[] — ordered inventory of every document in the submission.dataFiles[] — companion structured data files; usually empty for S-3.seriesAndClassesContractsInformation[] — fund/series/class metadata, populated only when the registrant is an investment company.entities[] — registrant and co-registrant blocks.id — an internal 32-character hash identifier for the record.Each documentFormatFiles[] entry carries sequence, size, documentUrl (the canonical sec.gov/Archives/edgar/data/... URL), description, and type. The type values map directly to EDGAR's exhibit taxonomy and recur across the dataset:
S-3 / S-3/A for the primary form;EX-1.x for the underwriting agreement or form of underwriting agreement;EX-3.x for charter and bylaws (typically incorporated by reference rather than re-filed);EX-4.x for instruments such as forms of certificate, indentures, supplemental indentures, and forms of notes or warrants;EX-5.x for legal opinions on the validity of the securities;EX-8.x for tax opinions;EX-12 for computations of ratios of earnings to fixed charges (largely historical after 2018);EX-23.x for consents of independent registered public accounting firms and other named experts;EX-24.x for powers of attorney;EX-25.x for the Form T-1 statement of trustee eligibility;EX-99.x for supplemental materials;EX-FILING FEES for the Exhibit 107 fee calculation table introduced in 2022;GRAPHIC for referenced image files (catalogued in metadata but not packaged in the ZIP).Each entities[] block describes one registrant party. Useful keys include companyName (suffixed with the role, e.g. (Filer), (Co-Registrant)), cik, irsNo, fileNo (the 333- Securities Act file number assigned to the registration), filmNo (EDGAR film number for that acceptance), act ("33" for the Securities Act), type (the form type for that party), sic (with both code and label), stateOfIncorporation, fiscalYearEnd (MMDD), and tickers[]. Multi-registrant shelf filings — common when a parent registers debt to be issued by a financing subsidiary, or when a holding company files alongside guarantors — produce one block per CIK, each tagged with its own role.
A Form S-3 registration statement is internally organized into the following ordered layers, all of which appear inside the primary HTML document unless otherwise noted.
The opening pages contain EDGAR's facing page, which carries the form name, the SEC file number assignment, the exact registrant name as incorporated, the state of incorporation, the IRS employer identification number, the SIC code, the principal executive office address and telephone number, and the name and address of the agent for service of process. Co-registrants, when present, appear in a tabular block on the same facing page with parallel columns of identifying information. The facing page also identifies whether the filing is a new registration, a post-effective amendment, or an automatic shelf, and contains the box checks that classify the registrant (large accelerated filer, accelerated filer, non-accelerated filer, smaller reporting company, emerging growth company) and indicate whether the registrant has elected to use the extended transition period for new accounting standards.
Historically, the facing page also carried a calculation-of-registration-fee table listing each title of securities being registered, the amount, the proposed maximum offering price per unit, the proposed maximum aggregate offering price, and the registration fee. Following the SEC's 2022 fee-disclosure modernization, that table moved out of the cover into the dedicated Exhibit 107 (EX-FILING FEES) attachment.
S-3/A amendments and many shelf filings include an explanatory note immediately after the cover that delimits the scope of the amendment, identifies the prior file number, references prospectus supplements filed under the same shelf, or describes the special structure of the offering (for example, a combined primary/secondary shelf, or a shelf with separate base prospectuses for different security types).
The prospectus is the substantive disclosure document and constitutes the bulk of the primary S-3. It opens with a prospectus cover page restating the offering's nature, the title and amount of securities, the manner of offering, and the trading symbol if applicable, followed by:
The prospectus typically closes with information about how to obtain incorporated documents and any required disclosures concerning the SEC's indemnification position.
Following the prospectus, every S-3 contains a Part II section. Part II includes Item 14 (other expenses of issuance and distribution), Item 15 (indemnification of directors and officers, summarizing state-law authority and charter/bylaw indemnity provisions), Item 16 (a list of exhibits, with each exhibit identified by number, description, and a cross-reference indicating whether it is filed herewith or incorporated by reference from a prior filing), and Item 17 (undertakings, including the standard Rule 415 shelf undertakings, the indemnification undertaking, and any form-specific undertakings such as those for offerings of asset-backed securities or for delayed offerings under Rule 415(a)(1)(x)).
Part II is followed by signature pages signed pursuant to the Securities Act by the registrant (through a duly authorized officer), the principal executive officer, the principal financial officer, the principal accounting officer, and a majority of the board of directors. Powers of attorney, when used, are referenced in the signature block and filed as Exhibit 24. Multi-registrant filings carry parallel signature blocks for each co-registrant.
The exhibit index in Part II numerically lists every exhibit. The actual exhibit documents are filed as separate components of the EDGAR submission and appear in the dataset record as separate HTML (or, occasionally, PDF) files, each wrapped in its own SGML envelope with the appropriate <TYPE>. Recurring exhibit roles include:
The canonical minimum exhibit set on a routine post-2022 S-3 shelf is primary form + EX-5.x legal opinion + EX-23.x auditor consent + EX-FILING FEES. Larger shelves and debt registrations expand the exhibit list considerably; S-3/A amendments may strip the exhibit set down to refreshed consents and a re-dated legal opinion when the amendment is purely procedural.
Each record packages, for one accession:
metadata.json JSON object;The full SGML .txt concatenation of the submission is referenced by metadata.json (linkToTxt) but is not shipped as a separate file inside the folder; the constituent documents already provide its content in extractable form.
.jpg, .gif, .png) — referenced in the submission's HTML and catalogued as GRAPHIC rows in documentFormatFiles[], but intentionally not packaged in the ZIP. Logos, signature scans, and chart images therefore appear as broken <img> references in the rendered HTML.S-3/A or POSASR) and produce separate dataset records..hdr.sgml header files — per-document header metadata is reflected in metadata.json rather than shipped as a separate file.Form S-3's anatomy has been stable in broad outline since the 1990s, but several material changes affect what one expects to find inside records of different vintages.
<TYPE> segments inside the SGML submission, with minimal HTML tagging. Calculation-of-registration-fee tables sat on the facing page. Exhibit 12 (ratio of earnings to fixed charges) was routinely filed by many registrants.EX-FILING FEES) presented in inline XBRL. From mid-2022 forward, S-3 records routinely include an EX-FILING FEES document carrying the fee calculation table; older records carry the same content embedded in the cover page instead.entities[] array in metadata.json reflects this directly.<TEXT> block of the primary document contained plain text with monospace tabular alignment; exhibit documents were similarly text-only. Records of this vintage preserve those text bodies inside their original SGML wrappers.<TEXT> block of each shipped document is HTML, and the rendered filings carry tables, fonts, and embedded image references. The dataset preserves these documents as .htm files with the EDGAR SGML header lines retained verbatim above the <HTML> root.EX-FILING FEES HTML document. Records from mid-2022 onward therefore include iXBRL fact tagging within that exhibit even though the surrounding prospectus remains plain HTML.S-3/A accession with the prior S-3 accession (and any earlier S-3/A accessions) sharing the same fileNo (333- number) on the registrant's entities[] block.<HTML> root must skip the SGML header lines (<DOCUMENT>, <TYPE>, <SEQUENCE>, <FILENAME>, <DESCRIPTION>, <TEXT>) and the closing </TEXT></DOCUMENT> tail. Conversely, parsers that key off the EDGAR <TYPE> value can use those header lines as a reliable label without re-parsing the HTML.entities[] element will miss co-registrants on shelf debt filings; the type field within each entity block distinguishes filer from co-registrants.<img> references to .jpg/.gif files that are catalogued in documentFormatFiles[] as GRAPHIC but are not present in the folder by design; this is expected and not a packaging defect.EX-5.1, EX-5.2 for two legal opinions covering different security tranches; EX-23.1 through EX-23.4 for separate audit and expert consents), and the same conceptual exhibit may carry different sub-numbers across filings. Classification by <TYPE> prefix (EX-5, EX-23, EX-FILING FEES) is more stable than the full sub-numbered token.seriesAndClassesContractsInformation[] carries fund/series/class identifiers that bind the S-3 to specific investment portfolios; for operating-company registrants, this array is empty and can be safely ignored.EX-5.x + EX-23.x + EX-FILING FEES is the canonical shape of a routine post-2022 S-3 shelf; deviations upward indicate debt instruments, multiple security tranches, or supplemental disclosures, while deviations downward typically indicate procedural amendments.Each record is a Form S-3 (initial filing) or Form S-3/A (amendment) submitted to EDGAR by a U.S. domestic issuer registering an offering of securities under the Securities Act of 1933. The filer is always the registrant whose securities are being registered, signed by its principal executive, financial, and accounting officers and a majority of its board on the Securities Act signature page. Where debt is issued with subsidiary guarantees, each guarantor co-registrant also signs and appears on the EDGAR header, so a single accession number can map to multiple CIKs.
Selling securityholders named in a secondary offering are not filers. They are listed in the prospectus and are statutory underwriters under Section 2(a)(11), but the registration statement itself is filed by the issuer and the dataset assigns it to the issuer's CIK.
A registrant must satisfy both the registrant tests in General Instruction I.A and at least one transaction test in General Instruction I.B.
Registrant requirements (I.A):
A late or missed 10-K, 10-Q, or specified 8-K item within the look-back period generally disqualifies the issuer until a fresh twelve months of timely reporting accrues. Loss of S-3 eligibility forces the issuer back to Form S-1.
Transaction categories (I.B):
A well-known seasoned issuer (Rule 405) meets I.A and either holds USD 700 million of non-affiliate common equity worldwide or has issued USD 1 billion of non-convertible securities for cash in registered primary offerings over the prior three years. Asset-backed issuers, registered investment companies, and recently disqualified bad actors are excluded. WKSIs may file an automatic shelf registration statement (ASR) on Form S-3 that is effective on filing under Rule 462(e), add issuers, classes, or guarantors via automatically effective post-effective amendments, and pay registration fees pay-as-you-go at takedown.
Form S-3 has no periodic deadline. It is offering-driven: the registrant files when it intends to conduct a registered offering. Typical triggers include:
Section 5(a) requires effectiveness before any sale; Section 5(c) restricts pre-filing offers.
S-3/A records reflect amendments to a previously filed S-3, typically driven by:
Individual takedown pricing flows through Rule 424(b) prospectus supplements, not S-3/A amendments, and 424(b) filings are not part of this dataset.
The Form S-3 Files Dataset sits inside the Securities Act registration family. Its closest neighbors fall into three groups: other Securities Act registration forms (S-1, S-4, S-8, F-3, S-3ASR), the offering-execution filings that operate against an effective S-3 shelf (Rule 424(b) prospectuses and FWPs), and the Exchange Act periodic reports (10-K, 10-Q, 8-K) the S-3 incorporates by reference. A separate axis is delivery format: this is a file-level package dataset, not a search index or a structured-fact API.
Used by issuers that do not qualify for S-3 short-form treatment, primarily IPOs and seasoned issuers lacking sufficient public float or reporting history.
Key differences:
Registers securities issued as consideration in mergers, exchange offers, and reorganizations.
Key differences:
Registers securities offered under employee benefit plans (options, ESPPs, RSUs, 401(k) employer-stock funds).
Key differences:
Direct foreign private issuer counterpart to S-3. Structurally near-identical (securities description, plan of distribution, incorporation by reference, exhibits).
Key differences:
Variant available to well-known seasoned issuers (WKSIs); effective immediately on filing, with pay-as-you-go fees and post-effective additions.
Key differences:
S-3 and S-3/A, so S-3ASR filings are not included. Total WKSI shelf activity requires pairing this dataset with an S-3ASR-scoped dataset.The most important distinction for anyone studying actual capital raised under a shelf.
Key differences:
Written marketing communications used during a registered offering that fall outside the statutory prospectus.
Key differences:
S-3 leans on these by reference: 10-K supplies audited financials, MD&A, business, and risk factors; 10-Q supplies interim financials; 8-K supplies material-event updates that often keep a shelf prospectus current.
Key differences:
This dataset delivers the full EDGAR submission package per accession (metadata plus all documents, excluding images). Two adjacent delivery formats cover the same underlying filings differently:
The same file-level vs. search vs. structured distinction applies symmetrically to the sec-api S-1, S-4, S-8, F-3, 424, and FWP file-level datasets: each delivers the complete EDGAR submission for its form-type scope and is complementary to, not a replacement for, search and XBRL APIs.
The Form S-3 Files Dataset is the authoritative source for the base prospectus, shelf authorization, registration-stage exhibits (Exhibit 5 legal opinions, Exhibit 23 consents, Exhibit 1 underwriting agreements), S-3/A amendment history, and the incorporation-by-reference scaffolding linking a shelf to an issuer's Exchange Act stream. It does not substitute for: 424(b) (deal pricing and tranche terms), FWP (marketing communications), S-1 (self-contained disclosure for IPO and ineligible issuers), F-3 (FPI counterpart), S-3ASR (WKSI automatic shelves), S-4 or S-8 (different transaction types), or the underlying 10-K, 10-Q, and 8-K reports it references. For shelf and follow-on research on seasoned domestic issuers, this dataset is the anchor that the others attach to.
The Form S-3 Files Dataset supports a defined set of capital-markets, legal, research, and compliance workflows, with each role attaching to a different layer of the filing.
Disclosure counsel and capital-markets associates use the dataset as a precedent library. They benchmark cover-page elections (automatic shelf, primary vs. secondary, Rule 415 reliance), model plan-of-distribution language for shelf takedowns and ATM programs, and compare EX-5.1 legality opinions and EX-23.1 auditor consents across recent precedents in the same state of incorporation. Output feeds drafting, opinion-committee review, and market-practice memos for working groups.
ECM desks mine the cover page and Exhibit 107 filing-fee table to extract maximum aggregate offering price, registered classes, and amounts added under Rule 462(b) or carried forward under Rule 415(a)(6). Plan-of-distribution text identifies offering structure (firm-commitment, block, ATM, equity line, rights) and the named underwriter or sales agent. Output drives shelf-replenishment calendars, capacity dashboards, and origination pitches.
Sell-side analysts and fund PMs track potential dilution and overhang. New shelves and replenishments size maximum issuance; use-of-proceeds language tests management's stated capital plan; selling-stockholder tables flag insider, sponsor, or PIPE exits. Analysts also confirm which 10-K, 10-Q, and 8-K filings are incorporated by reference into the offering's information set. Output feeds dilution models, share-count waterfalls, borrow assumptions, and updated price targets.
For S-3 registrations covering debt, preferred, depositary shares, or hybrids, DCM teams read the description-of-securities sections, indenture exhibits incorporated by reference, and the fee table to estimate remaining shelf capacity by class. Used for takedown calendars, league-table prep, and credit-investor outreach ahead of benchmark deals.
Compliance staff at issuers, underwriters, and law firms verify continued S-3 eligibility. Focus is on Instruction I.A reporting history, I.B.1 public float, the WKSI checkbox, subsidiary co-registrant lists for guarantor structures, and the exhibit list (consents, legality opinions). Output supports periodic eligibility reviews, broker-dealer due diligence files, and post-effective amendment workflows.
M&A advisers use S-3 precedents to study how acquirers register shares for resale by target shareholders, handle earn-out shares, and fund cash consideration through shelf takedowns. Plan-of-distribution sections, selling-stockholder tables, and integration-related risk factors support precedent-transaction analysis and registration-rights drafting.
Researchers studying SEOs, shelf registration, and cost of capital build samples from cover-page metadata, fee tables, and plan-of-distribution text. The dataset supports event studies on announcement returns, WKSI take-up after 2005, ATM adoption trends, and text-based studies of disclosure tone around issuance.
Data engineers at market-data firms parse the cover page for issuer identifiers, security class, and offering size; extract Exhibit 107 for normalized fee and amount data; and tag plan-of-distribution text by offering type. Coverage from 1994 supports backfilled product launches and law-firm and audit-firm league tables built from EX-5.1 and EX-23.1.
Teams building retrieval-augmented systems and domain LLMs for capital markets use S-3 as a structured training and evaluation corpus. The repetitive base-prospectus, incorporation-by-reference, and exhibit structure supports fine-tuning extraction models for offering size, security type, underwriter identity, and shelf status, plus opinion- and consent-letter classifiers for document-review tooling.
Staff at financial regulators use cover-page data, fee tables, and S-3/A amendment history to study issuance patterns, monitor shelf take-up by industry, and assess registrant behavior against current eligibility rules. Output feeds trend reporting, rulemaking impact analysis, and supervisory review.
The Form S-3 Files Dataset supports a focused set of capital-markets, legal, and quantitative workflows that anchor on specific layers of the registration statement and its exhibits.
Parse the EX-FILING FEES document in each post-2022 record to extract registered security classes, amount registered, proposed maximum aggregate offering price, fee rate, and Rule 457(p)/415(a)(6) carry-forward offsets. Joining these rows to subsequent 424(b) takedowns yields a per-issuer dashboard of remaining shelf capacity by security type, used by ECM and DCM desks for replenishment calendars and origination pitches.
Filter documentFormatFiles[] by type prefix EX-5 and EX-23 to pull every legal opinion and auditor consent in the dataset. Group by counsel signatory and accounting firm to construct law-firm and audit-firm league tables, and use the opinion text as a precedent library for drafting validity language tied to state of incorporation and security type (common, preferred, debt, warrants).
Extract the selling-securityholder section of resale S-3 prospectuses to identify named PIPE investors, sponsors, and insiders, the share counts being registered for resale, and beneficial ownership before and after. Combined with cover-page issuer identifiers and filedAt, this feeds dilution models, share-count waterfalls, and overhang flags for sell-side coverage and long/short books.
Iterate entities[] across each S-3 to enumerate parent registrants, financing subsidiaries, and guarantor co-registrants, keyed by CIK and tagged with role ((Filer), (Co-Registrant)). Joining with EX-25 trustee qualifications and EX-4 indenture references produces a structured map of guarantor chains used for credit analysis, collateral review, and corporate-tree maintenance.
Group records by registrant CIK and the 333- fileNo carried in entities[] to chain each base S-3 to its S-3/A amendments. Joining outward to the 424(b) and FWP datasets through the same file number reconstructs the full shelf lifecycle from authorization through individual takedowns, supporting capital-raise event studies and post-effective amendment review queues.
Segment the prospectus body to isolate the plan-of-distribution section across at-the-market programs, firm-commitment follow-ons, equity lines, rights offerings, and resale registrations. The resulting tagged corpus supports drafting checklists for capital-markets counsel and feeds fine-tuning sets for LLM-based offering-document classifiers and extractors.
Use cover-page box checks (large accelerated filer, accelerated filer, smaller reporting company, emerging growth company), the WKSI indication, and the SIC and state-of-incorporation fields in entities[] to track registrant eligibility profiles across vintages. Compliance teams use this to flag issuers approaching the I.B.1 public-float threshold or losing seasoned status; regulators use the same signals for rulemaking impact analysis on shelf and ATM activity.
The Form S-3 Files Dataset can be accessed in three ways: through the dataset index JSON API for metadata and container discovery, through the full dataset archive download, or through individual container ZIP files. The typical workflow is to poll the index endpoint to detect newly updated containers, then download only the changed containers to keep a local copy in sync.
Dataset Index JSON API: https://api.sec-api.io/datasets/form-s3-files.json
Returns dataset-level metadata (name, description, last updated timestamp, earliest sample date, total records and size, form types covered, container format, and file types) along with the full list of container files. Each container entry includes its key, size, record count, last updated timestamp, and direct download URL. This endpoint does not require an API key and is intended to be polled regularly to determine which containers were refreshed in the most recent update run, so that only changed containers need to be re-downloaded.
1
{
2
"datasetId": "1f13365b-9ae0-68ed-8c3d-4ccca853f937",
3
"datasetDownloadUrl": "https://api.sec-api.io/datasets/form-s3-files.zip",
4
"name": "Form S-3 Files Dataset",
5
"description": "Form S-3 is a short-form registration statement used to register securities offerings under the Securities Act of 1933.",
6
"updatedAt": "2026-05-06T02:50:50.342Z",
7
"earliestSampleDate": "1994-01-01",
8
"totalRecords": 318556,
9
"totalSize": 6390772596,
10
"formTypes": ["S-3", "S-3/A"],
11
"containerFormat": "ZIP",
12
"fileTypes": ["TXT", "JSON", "HTML", "PDF"],
13
"containers": [
14
{
15
"downloadUrl": "https://api.sec-api.io/datasets/form-s3-files/2026/2026-05.zip",
16
"key": "2026/2026-05.zip",
17
"size": 13818783,
18
"records": 154,
19
"updatedAt": "2026-05-06T02:50:50.342Z"
20
}
21
]
22
}
Download Entire Dataset: https://api.sec-api.io/datasets/form-s3-files.zip?token=YOUR_API_KEY
Downloads the complete Form S-3 Files Dataset as a single ZIP archive containing all monthly containers from the earliest sample date (1994-01-01) to the latest refresh. This endpoint requires a valid API key passed via the token query parameter. Use this for the initial bulk load of the dataset.
Download Single Container: https://api.sec-api.io/datasets/form-s3-files/2026/2026-05.zip?token=YOUR_API_KEY
Downloads one individual monthly container ZIP file as referenced by the downloadUrl field in the index API response. This endpoint requires a valid API key. Use this for incremental updates after the initial bulk load, fetching only the containers flagged as updated by the index API.
The dataset covers Form S-3 (initial registration statements) and Form S-3/A (amendments) accepted by EDGAR. EDGAR's distinct S-3ASR form type for WKSI automatic shelf registrations is not included in this dataset's declared scope.
One record is a single Form S-3 or Form S-3/A registration statement at the granularity of one EDGAR accession number, packaged as a folder named with the 18-digit accession (dashes removed). The folder contains a metadata.json file plus the primary registration document and every non-image exhibit transmitted with the original EDGAR submission.
Form S-3 is filed by a U.S. domestic seasoned reporting issuer that satisfies the registrant tests in General Instruction I.A (twelve months of timely Exchange Act reporting, current on dividends and material indebtedness, U.S.-organized) and at least one transaction test in General Instruction I.B (most commonly the USD 75 million public-float threshold for primary offerings, the listed-class condition for secondary resales, or the WKSI thresholds for automatic shelves). Foreign private issuers cannot use S-3; they file F-3 instead.
The dataset begins at an earliest sample date of 1994-01-01, the start of comprehensive EDGAR coverage, and is refreshed forward on a monthly cadence. Containers are organized by year and month (for example, 2025/2025-06.zip).
Containers are distributed as monthly ZIP archives organized by year. Inside each container, individual filing components are preserved as HTML/HTM (the dominant format for the primary document and exhibits), JSON (the per-record metadata.json), TXT (legacy ASCII bodies inside the SGML envelope), and PDF (used for certain exhibit classes and permitted unofficial copies).
The S-3 establishes shelf capacity and the base prospectus; the 424(b) supplement executes each takedown and carries the actual deal pricing, tranche size, underwriters, and security-specific terms (coupon, maturity, conversion). For complete shelf-and-takedown lifecycle analysis, the S-3 record must be joined to the chain of 424(b) filings that reference its 333- file number.
Image binaries (.jpg, .gif, .png) are catalogued as GRAPHIC rows in documentFormatFiles[] but are intentionally excluded from the ZIP packaging. As a result, logos, signature scans, and chart images appear as broken <img> references when the HTML is rendered; this is by design, not a packaging defect.