The Form 424B3 Files Dataset is a complete corpus of prospectuses filed on EDGAR under Rule 424(b)(3) of the Securities Act of 1933 — the rule that governs substantive, non-pricing updates to a prospectus previously filed as part of an already-effective registration statement. A single record is one EDGAR submission, materialized on disk as one accession-numbered folder containing a structured metadata.json, the primary prospectus document (almost always .htm), and any accompanying exhibits (most commonly an EX-FILING FEES inline-XBRL table, intermittently EX-4.x, EX-10.x, and EX-99.x exhibits). The underlying filings are submitted to EDGAR by the Securities Act registrant named on the prospectus cover, spanning domestic operating companies on Form S-1/S-3, foreign private issuers on F-1/F-3/F-4, banks and structured-products issuers, business development companies, asset-backed securities depositors, SPACs, and smaller reporting companies. Coverage runs from January 1, 1994 — when EDGAR filing became mandatory for most domestic registrants — through the current monthly refresh. Monthly container ZIPs contain TXT, JSON, HTML, and PDF files.
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 captures every Form 424B3 submission accepted by EDGAR from January 1, 1994 to the present. Form 424B3 is the EDGAR form tag applied when an issuer files a final prospectus under Rule 424(b)(3) that reflects substantive changes from or additions to a prospectus previously filed as part of an effective registration statement (typically Form S-1, S-3, S-4, Form F-1, F-3, or F-4). It is not a stand-alone registration; it is the mechanism by which a registrant refreshes the currently-effective prospectus between post-effective amendments. Typical uses include incorporating updated financial statements from newly filed Exchange Act reports, reflecting transactions such as selling-stockholder additions or private-placement resale registrations, delivering pricing or terms for a takedown off a shelf, or attaching a Form 8-K whose contents must be delivered to purchasers. The form therefore sits at the intersection of the registration-statement world (its legal authority) and the ongoing-reporting world (the content it frequently incorporates).
Within the 424(b) family, 424B3 is specifically reserved for substantive prospectus updates, distinct from 424B1 (information omitted in reliance on Rule 430A), Form 424B2 (primary-offering pricing), Form 424B4 (Rule 430A final prospectus), and Form 424B5 (Rule 415 shelf takedowns), although filer use of the boundaries between these subtypes is not entirely uniform in practice.
The dataset is distributed as monthly container ZIPs under a {year}/{year}-{month}.zip path convention; inside each container, one folder per accession holds the metadata envelope and every document EDGAR received for that submission except image files. Container file types are TXT, JSON, HTML, and PDF.
A single record in the Form 424B3 Files Dataset is one EDGAR submission accepted under Rule 424(b)(3) of the Securities Act of 1933, materialized on disk as one accession-numbered folder. Folders sit flat under a monthly directory produced by unpacking {year}/{year}-{month}.zip into {year}-{month}/{18-digit-accession}/; the folder name is the EDGAR accession number with dashes stripped (for example 0001839882-25-062642 becomes 000183988225062642). No intermediate grouping by filer, CIK, or day is used.
Each folder contains exactly one metadata.json describing the submission, one primary prospectus document (almost always .htm, extremely rarely with an accompanying .pdf), and zero or more exhibit documents that travelled with the prospectus through EDGAR. The record unit is the accession, not the individual document: a 424B3 filing that ships with several exhibits is still one record, whose folder contains several files.
In the reference container (October 2025, 569 accession folders, 1,207 files), 505 folders held exactly two files (metadata plus one primary .htm), 61 held three, 2 held four, and a single outlier held six (a Galaxy Digital filing with EX-4.1, EX-10.1, EX-99.1, and EX-99.2). Primary-document HTML sizes span roughly 2.7 KB to 15.5 MB, with a median near 37 KB and mean near 386 KB — the mean is dragged up by a thin tail of full prospectus supplements.
Each record folder combines a structured metadata envelope with the set of document files from the original EDGAR submission:
metadata.json — a JSON object derived from the EDGAR SGML submission header, describing the accession, the filer(s), and every document EDGAR received.<TYPE>424B3 inside the SGML <DOCUMENT> wrapper, containing the prospectus body itself.EX-FILING FEES inline-XBRL table, and less commonly EX-4.x, EX-10.x, and EX-99.x exhibits..pdf rendering of the prospectus, extremely rarely (exactly one occurrence among 569 folders in the October 2025 reference month).Image assets enumerated on EDGAR as GRAPHIC documents (company logos, signature scans, chart images) are deliberately omitted from the folder even though they remain listed in metadata.json.documentFormatFiles, consistent with the dataset's "all documents except image files" rule. Extracted XBRL instance XML sidecars referenced in dataFiles are similarly not shipped.
Filenames are filer-agent conventions rather than EDGAR-standard metadata, but they cluster per filer and carry useful routing information:
jpm_424b3-{termsheet-id}.htm.ea{doc-id}-01_424b3.htm or ea{doc-id}-424b3_{ticker}.htm.d{id}d424b3.htm or 424b3_{issuer_slug}.htm.tm{id}d{seq}_424b3.htm paired with tm{id}d{seq}_ex-filingfees.htm.ps10371.htm, form424b3.htm, listing_{YYYYMMDD-HHMM}.htm, sales_{YYYYMMDD-HHMM}.htm, kittprosupp1031252.htm, usq424b3liquidationnotices.htm. Timestamps baked into filenames for marketplace-lending listing and sales reports are filer conventions and reconstruct intra-day filing cadence for those filers.Every document shipped in the record preserves the SEC SGML envelope that wraps the actual HTML payload. The envelope carries four tag lines before the body and terminates the document after it:
1
<DOCUMENT>
2
<TYPE>424B3
3
<SEQUENCE>1
4
<FILENAME>jpm_424b3-36350.htm
5
<DESCRIPTION>SUMMARY OF TERMS
6
<TEXT>
7
<HTML>
8
... the prospectus body ...
9
</HTML>
10
</TEXT>
11
</DOCUMENT>
<TYPE> carries the EDGAR document-type code: 424B3 for the primary document, and EX-FILING FEES, EX-4.1, EX-10.1, EX-99.1, EX-99.2, GRAPHIC, XML, etc. for exhibits and sidecars.<SEQUENCE> is the ordinal position of the document within the submission.<FILENAME> matches the on-disk filename inside the record folder.<DESCRIPTION> is the filer-supplied human description shown on the EDGAR filing index (for example SUMMARY OF TERMS, EX-FILING FEES, PRESS RELEASE, GRAPHIC).<TEXT> opens the payload block containing the HTML document itself.HTML parsers that require a clean document root should strip this envelope or begin parsing at <HTML>. The envelope is the authoritative source for <TYPE> and <DESCRIPTION>; <FILENAME> is the exact on-disk name used to join the envelope back to metadata.json.documentFormatFiles.
The primary document is the prospectus supplement itself. Its internal content varies sharply with the kind of offering being updated, and the dataset spans the full range of 424B3 use:
ps10371.htm-style documents). These one- or two-page supplements state trade date, original issue date, maturity date, principal amount, interest rate and frequency, redemption provisions, CUSIP, the agent's capacity, and typically close with a short validity opinion of counsel.listing_20251030-1500.htm and matching sales_… files). These are machine-generated HTML tables enumerating individual Borrower Payment Dependent Notes series with loan amount, Prosper Rating, term, borrower yield, and the corresponding note series identifiers; they recur many times per day for a single filer and account for a substantial share of total record volume.Across all variants the document carries issuer identification in its header, references the underlying registration-statement file number (the 333- number to which the 424B3 relates), and concludes with delivery/date information indicating when the supplement is being delivered to investors.
Exhibits, when present, sit alongside the primary document in the same folder. Observed exhibit types:
EX-FILING FEES — an inline-XBRL HTML document implementing the Calculation of Filing Fee Table for Rule 424(b)(3). It uses the SEC ffd: taxonomy (https://xbrl.sec.gov/ffd/2025/ffd-2025.xsd) and tags facts such as ffd:AmtSctiesRegd, ffd:MaxAggtOfferingPric, ffd:FeeRate, ffd:FeeAmt, ffd:RegnFileNb, and dei:EntityCentralIndexKey. When this exhibit is present, metadata.json.dataFiles typically references an extracted XBRL instance XML sidecar, which is itself not shipped inside the folder.EX-4.x — forms of securities, indentures, supplemental indentures, or warrant/note agreements relevant to the offering.EX-10.x — material contracts referenced by or incorporated into the prospectus, such as securities-purchase agreements or registration-rights agreements.EX-99.x — press releases, investor presentations, consent letters, or other supporting materials. In some records these exhibits drive the substantive update (for example an EX-99.1 carrying financial statements of an acquired business).In the October 2025 reference month, roughly 89% of accessions ship with just metadata plus the primary document, roughly 11% ship one additional exhibit (overwhelmingly the EX-FILING FEES iXBRL file), and a small residual ship multi-exhibit bundles.
metadata.json anatomyThe metadata object is the structured lens over the filing. Observed top-level fields:
formType — always the string "424B3" for this dataset.accessionNo — canonical dashed EDGAR accession (for example "0001839882-25-062642").description — EDGAR form description, typically "Form 424B3 - Prospectus [Rule 424(b)(3)]".filedAt — ISO-8601 acceptance timestamp with timezone offset.id — 32-character hex dataset identifier.linkToFilingDetails — URL of the primary document under www.sec.gov/Archives/edgar/data/....linkToTxt — URL of the complete submission .txt (SGML bundle) on EDGAR.linkToHtml — URL of the EDGAR filing-index page for the accession.linkToXbrl — URL of a stand-alone XBRL instance if one exists; frequently empty for 424B3.documentFormatFiles — array describing every document in the EDGAR submission. Each entry carries sequence (ordinal, space for the aggregate .txt), size (bytes as a string), documentUrl, description (the SGML <DESCRIPTION> value), and type (the SGML <TYPE> value). This array includes GRAPHIC entries for images that exist on EDGAR but are deliberately excluded from the dataset folder.dataFiles — array with the same shape as documentFormatFiles, reserved for non-document data attachments such as the EXTRACTED XBRL INSTANCE DOCUMENT XML sidecar to the filing-fees exhibit. Frequently empty.seriesAndClassesContractsInformation — investment-company series/class identifiers; typically empty for 424B3 prospectuses.entities — one or more filer entities. Each object includes companyName (registered name with a role suffix such as (Filer)), cik, fileNo (Securities Act registration file number, e.g. 333-270004), irsNo, filmNo, type (form type as filed), act ("33" for Securities Act filings), stateOfIncorporation (two-letter code), fiscalYearEnd as MMDD, sic (SIC code plus label, e.g. "6021 National Commercial Banks"), and tickers (present for most operating-company filers; absent for finance-subsidiary issuing vehicles and other non-listed entities).Co-registrant filings (for example an operating company plus a finance-subsidiary issuer, or a marketplace-lending platform plus its funding vehicle) produce multiple entity objects — each with its own CIK, file number, and film number — which is the structural signal that more than one legal entity is on the filing. The 333- file number in entities[*].fileNo is the bridge from the 424B3 to the underlying effective registration statement; that registration statement is structurally separate and is not shipped in this dataset.
The record folder contains:
metadata.json for the accession.424B3 in the submission (the primary prospectus).EX-* exhibit document (most commonly EX-FILING FEES; intermittently EX-4.x, EX-10.x, EX-99.x).<TYPE>, <SEQUENCE>, <FILENAME>, <DESCRIPTION>, <TEXT>, and <HTML> markers.Referenced by metadata.json but not shipped inside the record folder:
GRAPHIC documents (image assets such as .jpg logos, signature scans, chart images, embedded figures) — excluded per the dataset's image-exclusion rule. Prospectus HTML may therefore contain <img> tags whose src filenames are absent on disk; those filenames remain visible in documentFormatFiles.*_htm.xml companion to the EX-FILING FEES exhibit) — typically listed in dataFiles but not materialized in the folder. The inline-XBRL facts themselves are still present inside the EX-FILING FEES .htm document..txt bundle — referenced via linkToTxt but not included.The form tag 424B3 has been continuously available on EDGAR from January 1994, but the substantive content and adjacent disclosure obligations have shifted:
EX-FILING FEES iXBRL exhibit using the ffd: taxonomy — a structural change visibly baked into the dataset, with filing-fee exhibits appearing in roughly one in ten records in a representative recent month.The on-disk format of the shipped documents tracks the broader EDGAR presentation timeline:
<DOCUMENT> envelope, with tables rendered in fixed-width spacing..htm file inside the SGML envelope, sometimes with embedded images that the dataset excludes.EX-FILING FEES exhibit. The 424B3 prospectus body itself is not iXBRL-tagged — 424B3 is not subject to the financial-statement iXBRL mandate — but the filing-fee exhibit is, and the SEC ffd: taxonomy facts can be consumed directly as structured data.metadata.json, then iterate documentFormatFiles to locate each payload by its on-disk <FILENAME>..htm file begins with the SGML header lines before <HTML>. Strip the envelope or start parsing at <HTML>; the envelope carries the authoritative <TYPE> and <DESCRIPTION>.<img> references in prospectus HTML whose filenames are missing on disk are the excluded GRAPHIC documents, not corruption. Extraction pipelines should tolerate absent image assets.entities signal multiple legal-entity identities on the filing (for example issuer plus guarantor, or operating company plus finance subsidiary). Each carries its own CIK, file number, and film number.fileNo is the bridge to the registration statement. The 333- file number in entities[*].fileNo links the 424B3 to the underlying S-1/S-3/S-4/F-1/F-3/F-4 needed to place the supplement in legal context. The base registration is not shipped in this dataset.The filer is always the Securities Act registrant named on the cover of the prospectus. Form 424B3 is submitted to EDGAR by the issuer that has an effective registration statement on file and needs to put on the record a prospectus, prospectus supplement, or revised prospectus that reflects a substantive, non-pricing change from, or addition to, information in a prospectus previously filed (or deemed filed) as part of that registration statement.
Filing parties are drawn from essentially every class of registrant that conducts registered offerings:
Underwriters, placement agents, and selling securityholders appear in the prospectus but do not file it. In multi-registrant structures (parent guarantor plus finance subsidiary, co-issuers, subsidiary guarantors), multiple CIKs may be associated with the submission, but the registrant on the cover is the operational filer.
A 424B3 is triggered whenever substantive, non-pricing information in a live prospectus must be changed, added, or refreshed. Typical triggers:
Rule 424(b)(3) requires the prospectus to be filed no later than the second business day following the earlier of (i) the date it is first used or (ii) the date of effectiveness of the post-effective amendment to which it relates, or transmitted by a means reasonably calculated to result in filing with the Commission by that date. A 424B3 is not subject to staff review or acceleration; it operates as a filed update under an already-effective registration statement.
Rule 424 governs prospectus filings made after a registration statement becomes effective. Rule 424(b)(3) is the paragraph for substantive, non-pricing changes to the last filed prospectus. The 424(b) paragraphs that commonly sit next to it:
Rule 415 authorizes delayed and continuous shelf registration offerings; Rules 430B and Rule 430C define what information may be omitted from a prospectus at effectiveness and later supplied through a 424(b) filing. Together with Section 10(a) of the Securities Act, these rules create the framework that 424(b)(3) operationalizes: keeping the Section 10(a) prospectus current between effectiveness and sale.
The modern 424(b)(1)-(b)(8) structure reflects the 1982 integrated disclosure reforms that introduced Rule 415 shelf registration and the 2005 Securities Offering Reform. EDGAR became mandatory for most domestic registrants between 1993 and 1996; the dataset's earliest records begin January 1, 1994. Pre-EDGAR paper 424(b)(3) prospectuses exist in the Commission's historical archives but are outside this dataset.
Form 424B3 belongs to the Rule 424(b) prospectus family, where each paragraph of the rule produces a distinct filing with its own trigger and content. The most useful comparisons are the sibling 424B-series filings, the base registration statements they supplement, Form 424H, free writing prospectuses, Rule 497/Rule 497K investment-company prospectuses, and post-effective amendments. All of these circulate around the same broad purpose — delivering current prospectus disclosure in a registered offering — but their triggers, scope, and legal mechanics diverge sharply.
Filed when a prospectus previously omitted information under Rule 430A (typically final pricing for an IPO) and is now complete. Form 424B1 is a pricing-completion vehicle; 424B3 is an update vehicle for substantive disclosure changes unrelated to 430A omitted information.
Used for primary shelf offerings under Rule 415 at market or non-fixed prices. A 424B2 is tied to a specific sale event and carries pricing/terms for that takedown. A 424B3 is not a takedown instrument and is not triggered by a sale; it carries substantive changes to previously filed prospectus disclosure.
Like 424B1, 424B4 supplies information previously omitted under Rule 430A, but is filed where the prospectus was otherwise complete at effectiveness. 424B3 never carries 430A pricing content.
The dominant vehicle for seasoned shelf takedowns, supplying information (issuer, selling holders, plan of distribution, security terms) left blank under Rule 430B. 424B5 is event-tied and Rule 430B–bound; 424B3 addresses substantive changes outside the 430B omitted-information framework and is not necessarily tied to a sale.
Narrowly scoped to Rule 430B(b) identification of selling security holders and resale amounts. 424B3 is broader and can update any substantive portion of the prospectus.
A periodic, time-driven refresh tied to the 16-month Section 10(a)(3) staleness clock, typically rolling in updated audited financials. Form 424B8 is event-driven, triggered by substantive disclosure changes rather than a statutory calendar.
The registration vehicles from which 424B3 prospectuses descend. The S/F-series defines what securities may be offered and is effective once; 424B3 is an ongoing update to the delivered prospectus within that effective registration. To reconstruct the original offering terms, go to the S/F filing; to see what purchasers actually received after substantive updates, go to the relevant 424B filings.
The preliminary prospectus for asset-backed securities under Regulation AB, filed before sale to allow investor review of pool-level disclosures. 424B3 is a final prospectus, not preliminary, and is not limited to ABS.
Written offering communications permitted under Rules 164 and 433, delivered alongside but outside the statutory prospectus. An FWP is not a Section 10 prospectus and cannot amend the delivered prospectus; a 424B3 is a statutory prospectus filing that modifies the prospectus itself.
Rule 497 prospectuses and Rule 497K summary prospectuses for registered investment companies (mutual funds, closed-end funds, variable products) filed under the Investment Company Act regime and Form N-1A/N-2. 424B3 sits in the parallel Rule 424(b) regime for operating-company issuers. The two tracks are mutually exclusive.
A POS AM amends the registration statement itself and can reopen SEC review or require re-acceleration. 424B3 avoids that by leveraging Rule 424(b)(3) to reflect substantive changes at the prospectus level within the already-effective registration. The choice turns on whether the change is material enough to require re-amendment of the registration or can be carried through a prospectus-level update (often via incorporation by reference of subsequent Exchange Act reports).
Form 424B3's distinctive role is narrow and specific: it is the Rule 424(b)(3) vehicle for reflecting substantive changes or additions to a previously filed prospectus in an already-effective registered offering, without filing a post-effective amendment to the registration statement itself. Unlike 424B1 and 424B4, it carries no Rule 430A pricing mechanics. Unlike 424B2 and 424B5, it is not a shelf-takedown pricing instrument tied to a sale event or to Rule 430B omitted information. Unlike 424B7, it is not constrained to selling-security-holder identification. Unlike 424B8, it is not driven by the 10(a)(3) updating clock. Unlike 424H, it is neither asset-backed nor preliminary. Unlike POS AM, it does not amend the registration statement. Unlike FWP, it is a statutory Section 10 prospectus rather than a permitted supplemental communication. Unlike 497/497K, it sits in the operating-company prospectus regime, not the investment-company regime. Related 424B-series datasets and the base S/F registration datasets are complements, not substitutes: a complete view of a registered offering typically requires pairing the original S-1/S-3/F-1/F-3 with the specific 424B filings that supplied pricing (424B1/B4), takedown terms (424B2/B5), resale identifications (424B7), periodic updates (424B8), and substantive disclosure changes (424B3). The Form 424B3 Files Dataset isolates exactly the substantive-change layer of that stack across the full 1994-to-present EDGAR history.
Rule 424(b)(3) prospectus supplements carry the substantive updates to an already-effective prospectus: new risk factors, revised use-of-proceeds language, selling-securityholder tables, plan-of-distribution changes, and structured-note termsheets. The professions below each consume a specific slice of that record.
Disclosure counsel and capital markets associates use the corpus as a precedent library when drafting take-downs, PIPE resale registrations, and ATM programs. They pull the primary prospectus HTML from peer issuers to benchmark risk-factor language, use-of-proceeds paragraphs, and distribution mechanics, and rely on the metadata JSON to filter by issuer, date, and parent registration statement. Exhibits supply underwriting agreements, legality opinions, and note or warrant forms.
Origination and syndicate teams treat 424B3 filings as a live feed of competitor shelf take-downs, resale registrations, and continuous offerings. They extract offering size, gross spread, over-allotment, lock-ups, and underwriter lineup from the prospectus text and read the filing-fee iXBRL tables for structured offering-amount and security-type fields. Output: pitch comps, league-table pipelines, and fee-and-terms benchmarks.
Fundamental and event-driven analysts scan 424B3 filings to track capital-structure changes in near real time. Selling-securityholder tables flag overhang and future secondary supply; use-of-proceeds language indicates whether cash funds growth, refinancing, or M&A; refreshed risk factors flag what the issuer now deems material. The prospectus plus metadata JSON feeds position-level dilution and overhang monitors.
Desks trading autocallables, buffered notes, market-linked CDs, and leveraged upside notes harvest pricing supplements filed as 424B3 to extract underlyings, participation rates, caps, buffers, barriers, call schedules, and estimated initial values. Filing-fee iXBRL records give aggregate program size. The parsed termsheets feed secondary-market pricing models, competitor payoff clustering, and periodic market-color on issuance volume and feature drift.
Compliance and surveillance teams at underwriters and distributors use metadata JSON fields (accession number, filing date, issuer CIK, parent registration form) to confirm timely Section 10(a)(3) refreshes, prospectus-delivery obligations, and supplement filing windows. Prospectus text supports content checks such as verifying that a supplement incorporates the triggering 10-K or 10-Q by reference.
Credit analysts and counterparty-risk groups read 424B3 filings as an early signal of changes to an obligor's capital stack. Security rank, covenant references, use of proceeds, and coupon or pricing mechanics drive re-runs of leverage and coverage models and updates to internal issuer files. For bank-issued notes, the filings quantify unsecured issuance for wholesale-funding analysis.
Short sellers and activist research shops mine resale registrations for shares held by PIPE investors, convertible holders, warrant holders, and consultants. Selling-securityholder tables, beneficial ownership footnotes, and plan-of-distribution language size the potential overhang; the gap between original financing and resale registration supports dilution-trajectory work and toxic-financing detection.
Quant researchers and NLP engineers use the corpus as a domain-specific training and evaluation set. Typical tasks: fine-tuning on risk-factor prose, classifying primary versus resale offerings, extracting offering size and underwriter identity, parsing structured-note payoff parameters, and building RAG indexes that cite back to the filed HTML. The ZIP-packaged metadata plus HTML and PDF supports both batch ingestion and per-filing lookup.
Finance and law academics study offering-timing patterns, prospectus readability, risk-factor evolution, structured-note design and pricing, and PIPE resale economics. Coverage from 1994 to present, with exhibits and metadata retained, supports long-horizon panels linking prospectus characteristics to issuer outcomes and regime changes.
Platform engineers ingest the dataset as a ready corpus instead of crawling EDGAR. Metadata JSON supplies the join keys to base registration statements, issuer masters, and downstream capital-structure tables. HTML and PDF feed full-text indexes and embedding stores; filing-fee iXBRL normalizes into offering-size tables. The result underlies internal research portals, surveillance tools, and RAG systems.
Lawyers pull precedent prose, bankers pull offering terms and iXBRL fee data, analysts pull selling-securityholder and use-of-proceeds tables, structured-products desks pull termsheet parameters, compliance pulls metadata timestamps, and NLP teams pull the full corpus. The common need is the complete 424B3 population with exhibits and metadata in one place.
Concrete workflows the Form 424B3 Files Dataset supports, each tied to specific record components.
Parse the Selling Stockholders table inside full prospectus supplements (DFIN-produced resale registrations are typical) to extract holder name, shares beneficially owned before, shares being registered, and shares owned after. Join across accessions on entities[*].cik to build a per-issuer running overhang count. Output feeds dilution and short-interest models and flags PIPE investors whose shares have just become tradable.
Iterate high-frequency issuer folders (e.g. jpm_424b3-{termsheet-id}.htm) and extract the summary-of-terms block: underlying index or basket, contingent-interest rate, trigger/call levels, observation schedule, CUSIP/ISIN, pricing date, and agent's discount. The parsed parameters feed autocallable payoff models, competitor feature-drift dashboards, and issuance-volume market color. The compact size (10-30 KB HTML) makes batch extraction cheap.
Load the EX-FILING FEES exhibit (present in roughly 11% of October 2025 records) and pull ffd:AmtSctiesRegd, ffd:MaxAggtOfferingPric, ffd:FeeRate, ffd:FeeAmt, and ffd:RegnFileNb from the inline-XBRL facts. Group by filer CIK and SIC to build primary-vs-resale league tables, fee-rate time series, and program-level cumulative-issuance tallies without scraping prospectus prose.
Index the primary 424B3 HTML (after stripping the SGML envelope) into a full-text and embedding store, filtered via metadata.json on entities[*].sic, stateOfIncorporation, and fileNo prefix. Counsel drafting a PIPE resale registration or ATM supplement queries for peer risk-factor paragraphs, plan-of-distribution language, and use-of-proceeds boilerplate, then cites back to the filed accession.
fileNoUse entities[*].fileNo (the 333- number) as the join key from each 424B3 accession to the underlying S-1, S-3, S-4, F-1, F-3, or F-4 in a separate dataset. The resulting graph reconstructs the full post-effective update trail for a given shelf: original registration, subsequent 424B3 substantive updates, and parallel 424B2/B5 takedowns. Useful for reconstructing what purchasers received at any given date.
Group Prosper Funding and similar filers' listing_{YYYYMMDD-HHMM}.htm and sales_{YYYYMMDD-HHMM}.htm files by timestamped filename and parse the machine-generated HTML tables of Borrower Payment Dependent Notes. Output is an intraday series of loan amount, Prosper Rating, term, borrower yield, and the corresponding note series identifiers, supporting secondary-market analytics and origination-volume tracking for marketplace-lending platforms.
Full prospectus supplements frequently embed an incorporated Form 8-K body directly in the 424B3 HTML (for example a bridge-loan 8-K inside a TuHURA-style supplement). Detect these by scanning for Item numbers and 8-K boilerplate inside the prospectus body; extract the embedded event disclosure as a structured record, timestamped to filedAt. This surfaces material events that are legally part of the delivered prospectus and may not otherwise be picked up by 8-K-only pipelines.
Dataset Index JSON API: https://api.sec-api.io/datasets/form-424b3-files.json
This endpoint returns the dataset's metadata, including name, description, last updated timestamp, earliest sample date (1994-01-01), total records and total size, form types covered (424B3), container format (ZIP), and the file types inside each container (TXT, JSON, HTML, PDF). It also returns the download URL for the full dataset archive and the full list of monthly container files, each with its own size, record count, updated timestamp, and download URL. Polling this endpoint is the recommended way to monitor which containers changed during the most recent refresh so only updated archives are re-downloaded on a given day. This endpoint does not require an API key.
Example response:
1
{
2
"datasetId": "1f13365b-9ade-61da-963f-6ff8c7d90d93",
3
"datasetDownloadUrl": "https://api.sec-api.io/datasets/form-424b3-files.zip",
4
"name": "Form 424B3 Files Dataset",
5
"updatedAt": "2026-04-22T02:58:03.686Z",
6
"earliestSampleDate": "1994-01-01",
7
"totalRecords": 218070,
8
"totalSize": 18370779760,
9
"formTypes": ["424B3"],
10
"containerFormat": "ZIP",
11
"fileTypes": ["TXT", "JSON", "HTML", "PDF"],
12
"containers": [
13
{
14
"downloadUrl": "https://api.sec-api.io/datasets/form-424b3-files/2026/2026-04.zip",
15
"key": "2026/2026-04.zip",
16
"size": 13818783,
17
"records": 154,
18
"updatedAt": "2026-04-22T02:58:03.686Z"
19
}
20
]
21
}
Download Entire Dataset: https://api.sec-api.io/datasets/form-424b3-files.zip?token=YOUR_API_KEY
Downloads the complete dataset as a single ZIP archive covering all Form 424B3 filings from January 1994 to the present. Use this for a one-time bulk load of the full history. This endpoint requires an API key.
Download Single Container: https://api.sec-api.io/datasets/form-424b3-files/2026/2026-04.zip?token=YOUR_API_KEY
Downloads one monthly container ZIP instead of the full archive. Container paths follow the YYYY/YYYY-MM.zip pattern returned in the containers array of the index response and are the preferred unit for incremental daily syncs. This endpoint requires an API key.
The dataset covers Form 424B3 — prospectuses filed under Rule 424(b)(3) of the Securities Act of 1933 that reflect substantive, non-pricing changes from or additions to a prospectus previously filed as part of an effective registration statement. Every accession in the dataset carries the EDGAR form tag 424B3.
One record is one EDGAR submission accepted under Rule 424(b)(3), materialized on disk as one accession-numbered folder. The folder contains a metadata.json derived from the EDGAR SGML submission header, the primary prospectus HTML, and any accompanying exhibits (most commonly an EX-FILING FEES inline-XBRL table, occasionally EX-4.x, EX-10.x, or EX-99.x exhibits). Image assets listed as GRAPHIC documents are deliberately omitted.
The Securities Act registrant named on the cover of the prospectus files it. That population spans domestic operating-company issuers on Form S-1 or S-3, foreign private issuers on F-1/F-3/F-4, S-3 and F-3 shelf registrants and WKSIs, banks and structured-products issuers updating medium-term-note and structured-note programs, business development companies, asset-backed securities depositors under Regulation AB, and SPACs and smaller reporting companies running resale registrations. Underwriters, placement agents, and selling securityholders are named in the prospectus but do not file it.
Rule 424(b)(3) requires the prospectus to be filed no later than the second business day following the earlier of (i) the date it is first used or (ii) the date of effectiveness of the post-effective amendment to which it relates, or transmitted by a means reasonably calculated to result in filing by that date. A 424B3 is not subject to staff review or acceleration.
The dataset's earliest sample date is January 1, 1994 — aligned with EDGAR becoming mandatory for most domestic registrants — and coverage continues through the current monthly refresh. Pre-EDGAR paper 424(b)(3) prospectuses exist in the Commission's historical archives but are outside this dataset.
The dataset is distributed as ZIP containers, one per calendar month under a YYYY/YYYY-MM.zip path convention. Inside each container, file types are TXT, JSON, HTML, and PDF. The primary prospectus document is almost always .htm wrapped in an SGML <DOCUMENT> envelope; PDF renderings are extremely rare.
Form 424B2 and 424B5 are shelf-takedown pricing instruments tied to a specific sale event (424B2 for primary takedowns at market or variable prices, 424B5 for seasoned shelf takedowns supplying Rule 430B information). Form 424B3 is not a takedown or pricing vehicle — it is the Rule 424(b)(3) route for substantive, non-pricing changes to a previously filed prospectus. The Form 424B3 Files Dataset isolates that substantive-change layer across the full 1994-to-present EDGAR history.
Each filer entity in metadata.json.entities carries a fileNo field — the 333- Securities Act registration file number. That number is the bridge from the 424B3 accession to the underlying S-1, S-3, S-4, F-1, F-3, or F-4 registration statement. The base registration itself is structurally separate and is not shipped inside this dataset.