The Form 497K3B Files Dataset is a closed archive of mutual-fund "Profile" filings made on EDGAR by registered open-end management investment companies under Rule 497 and former Rule 498 of the Securities Act of 1933. Each record corresponds to one accepted EDGAR submission of Form 497K3B — a short, plain-English standardized disclosure document covering a fund's objective, principal strategy, risks, fees, performance, management, and purchase and redemption mechanics. The filer of record is always the trust, corporation, or other legal vehicle that holds the Form N-1A registration; a single record routinely covers multiple registered series and share classes within that registrant. The Profile regime was introduced by the SEC in 1998 (Release No. 33-7513) and discontinued effective March 30, 2009 by Release No. 33-8998, which adopted the Summary Prospectus framework as its replacement. The dataset spans October 1998 through March 2009, the entire operative lifetime of the form, and is delivered as monthly ZIP containers organized by year.
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 packages every accepted EDGAR submission of Form 497K3B over the form's entire active lifetime, from October 1998 through its discontinuation on March 30, 2009. Form 497K3B was the EDGAR submission type used to file a "Profile" prepared in accordance with former Rule 498 — specifically, profiles described in former Rule 497(k)(1)(iii)(B). The Profile, introduced by the SEC in 1998, was a short, plain-English, standardized disclosure document — typically a few pages of typeset HTML — that summarized the most decision-relevant attributes of a mutual fund. Investors receiving the Profile had the legal option of either purchasing fund shares directly on the basis of the Profile or first requesting the full statutory prospectus.
The Profile's content was prescribed by an itemised template requiring eleven question-form disclosure topics, presented in a fixed order, so each filing reads as a near-identical fill-in-the-blank document with issuer-specific narrative, fee tables, and performance data. The dataset covers the entire filer population that elected to use the Profile mechanism during 1998–2009; because Profile use was always optional, fund families that delivered only their full statutory prospectus during this window are simply absent from the corpus rather than missing from it. The form was rescinded effective March 30, 2009, when the SEC's Summary Prospectus framework supplanted the original Profile regime, so no filings of this type exist after that date. Records are distributed as monthly ZIP containers, with file types limited to TXT, JSON, and HTML.
A single record in the Form 497K3B Files Dataset corresponds to one accepted EDGAR submission of Form 497K3B — that is, one mutual-fund "Profile" filing made by an open-end management investment company under Rule 497 of the Securities Act of 1933 and the Profile framework of former Rule 498. Each record is materialised as a folder named with the 18-digit accession number (no dashes), e.g. 000031321207000114/, residing inside a monthly partition of the form <YYYY>/<YYYY-MM>.zip. Every folder pairs one machine-readable header file (metadata.json) with the original primary document submitted to EDGAR for that accession. The unit of analysis is the accession, not the fund: a single 497K3B record routinely covers multiple registered series, multiple share classes, and — for trust-style filers — several sister funds bundled into one Profile.
Each record folder contains:
metadata.json — a flat JSON header always present. It captures the EDGAR acceptance metadata, the registrant block, the series/class registry, and an inventory of every document in the original SGML submission.<TYPE>497K3B, with a filename chosen by the filer (e.g. isf.htm, inteq.htm, gen-ira.htm). The filename is not standardised and must be resolved through the documentFormatFiles inventory in metadata.json rather than assumed.A small number of records additionally include short ancillary text documents (e.g., transmittal letters), but the structural minimum is the metadata file plus the primary 497K3B HTML. Embedded raster artwork (bar charts, performance graphs, fund logos) is referenced from inside the HTML as <img src="iXXXXXXX.gif">, but the GIF binaries are intentionally excluded from the dataset — the file-types found in the dataset are TXT, JSON, and HTML, with image references in the HTML left as dangling pointers to the original EDGAR archive locations recorded in metadata.json. The dashed/canonical accession number (e.g., 0000313212-07-000114) is preserved inside the metadata file even though the folder itself uses the un-dashed 18-digit form.
metadata.jsonThe metadata file is a single JSON object with a flat top level plus a small number of structured arrays. The fields that carry intentional meaning are:
formType — always the string "497K3B".accessionNo — canonical dashed accession identifier, e.g. "0000313212-07-000114".filedAt — full ISO-8601 acceptance timestamp with timezone offset, e.g. "2007-11-19T08:19:48-05:00".effectivenessDate — calendar date on which the Profile took effect under Rule 497; present on filings that declare an effectiveness date and absent on others.description — boilerplate human label, typically "Form 497K3B - Profiles for certain open-end management investment companies, [Rule 497(k)(1)(iii)(B)](https://www.ecfr.gov/current/title-17/chapter-II/part-230/subject-group-ECFR503bf91e47b67cd/section-230.497)".linkToFilingDetails — direct URL to the primary 497K3B HTML document on EDGAR.linkToTxt — URL of the complete .txt SGML submission bundle on EDGAR (the canonical full filing).linkToHtml — URL of the EDGAR filing index page (*-index.htm).linkToXbrl — empty string for this dataset; 497K3B filings carry no XBRL exhibits.id — 32-character hexadecimal per-record fingerprint.documentFormatFiles[] — an inventory of every document in the original EDGAR submission. Each entry carries sequence (string ordinal, with the trailing complete-submission row using a single space), size (bytes as a string), documentUrl (absolute EDGAR archive URL), type ("497K3B" for the primary document, "GRAPHIC" for embedded GIF images, blank for the complete-submission text), and an optional description (e.g., "Complete submission text file"). Entries of type GRAPHIC are listed for completeness; their binaries are not packaged in the dataset.dataFiles[] — present but empty for this dataset (no XBRL/structured data files exist for 497K3B filings).entities[] — one or more registrant entries. Each entity carries companyName (suffixed with " (Filer)"), cik, irsNo, fileNo, filmNo, fiscalYearEnd (MMDD format), act ("33" for Securities Act of 1933 filings), type ("497K3B"), and optionally stateOfIncorporation. Multi-filer records are common — a single accession may aggregate several related funds filing under one transmission, in which case entities lists each as a separate object.seriesAndClassesContractsInformation[] — present when the filing identifies registered series and share classes under EDGAR's series/class identifier system. Each item contains series (the S00… series identifier), name (the series's marketed fund name), and classesContracts[] of { ticker, name, classContract } (the C00… class identifier and its trading symbol). Multi-series filings are routine for fund-family Profiles.Each primary document is wrapped in EDGAR's SGML document envelope inherited from the original submission, then carries a single mutual-fund Profile in HTML inside the <TEXT> block:
1
<DOCUMENT>
2
<TYPE>497K3B
3
<SEQUENCE>1
4
<FILENAME>isf.htm
5
<TEXT>
6
<html>
7
<body> ... fund profile content ... </body>
8
</html>
9
</TEXT>
10
</DOCUMENT>
The HTML inside is presentationally typeset rather than semantically structured. Practically every paragraph is a <div> carrying explicit text-align, margin-*, text-indent, and width attributes, with a nested <font face="Berkeley Book|Trajan|MetaPlusLF-…" size=… color=…> controlling typography. Section titles appear as bold runs of text inside divs rather than as <h1>/<h2> headings. Page headers and footers from the print layout survive as repeated bold Fund Profile strings and bare page digits between sections. Encoded EDGAR redline markers from prior amendments survive as literal <R> ... </R> brackets sprinkled through the HTML, marking text changed relative to a prior version of the Profile. Embedded images are referenced via <img src="iXXXXXXX.gif">, with the targets absent from the dataset.
Every 497K3B body follows the rule-prescribed Profile template, recognisable by a fixed sequence of question-form headings. A short prelude — effective date, the bold "Fund Profile" header, the fund name and ticker (e.g., "T. Rowe Price International Stock Fund — PRITX"), a one-line strategy tagline, and a disclaimer that the Profile summarizes information also contained in the statutory prospectus — precedes the eleven canonical sections, which appear in this order:
Multi-fund Profiles use plural variants of the same headings — for example, What are each fund's objectives and principal investment strategies? and How can I tell which fund is most appropriate for me? — and repeat the per-fund tables (performance, fees, expense example) side by side or stacked, one block per fund covered by the single accession.
The document closes with an adviser/distributor block giving the adviser's legal name and street address, the fund family's website, an internal piece-tracking code (e.g., RPS F37-035), and the legal name of the principal underwriter/distributor.
Each record contains the metadata header, the primary 497K3B HTML document with its full Profile narrative, embedded fee and performance tables, redline markers, all inline typography, and any short ancillary text documents that accompanied the original submission. The metadata file additionally preserves the full document inventory of the original SGML submission, the multi-filer entity block, the series/class identifier registry, and direct EDGAR URLs back to the canonical artifacts.
GIF graphics referenced from the Profile HTML — performance bar charts, fund logos, decorative rules — are deliberately omitted from the dataset, even though they are listed as GRAPHIC-type entries in documentFormatFiles[]. The complete-submission .txt SGML bundle from EDGAR is not packaged either; the metadata's linkToTxt points back to the original EDGAR copy. Statutory prospectuses, statements of additional information, shareholder reports, and other companion fund-family filings are filed under separate EDGAR form types (485APOS, 485BPOS, Form 497, N-1A, N-CSR, etc.) and are not part of a 497K3B record.
The 497K3B form had a comparatively short and stable life. It was introduced in 1998 with the SEC's adoption of Rule 498 ("Profiles for Certain Open-End Management Investment Companies"), which prescribed the eleven-question Profile content format described above; the dataset's earliest filings, dating from October 1998, already follow the canonical question template. Throughout the 1998–2009 window the headings, section sequence, and required tables remained materially constant: the Shareholder Fees / Annual Fund Operating Expenses pair, the Expense Example, and the Average Annual Total Returns table are stable structural fixtures. Incremental refinements over the period — for example, expanded after-tax return columns following the 2001 SEC amendments to standardized performance presentation, and progressively more detailed risk-factor enumeration as fund strategies grew more complex — manifest as additional rows or paragraphs within the existing sections rather than as new top-level headings. Adoption of EDGAR's series-and-class identifier system in 2006 caused seriesAndClassesContractsInformation to populate consistently for filings from that point onward; earlier records may lack this block. The form was rescinded by the SEC effective March 30, 2009, when the Summary Prospectus regime under amended Rule 498 (Release No. 33-8998) replaced the Profile framework with a similar but reorganised summary disclosure required at the front of the statutory prospectus and filed under the Form 497K and 497VPI form types; no 497K3B records exist after that date.
497K3B was an HTML-era form throughout its existence. EDGAR's mandatory acceptance of HTML had taken effect before the Profile regime was introduced, so even the earliest 1998 filings are HTML wrapped in the standard SGML <DOCUMENT> envelope rather than the plain-ASCII <TYPE>497K3B blocks characteristic of pre-1996 filings. Across the decade-long window the visual rendering moved from plainer HTML produced by word-processor exports toward the heavily inline-styled, font-tagged output of professional financial-typesetting software (the form seen in the late-period documents, with explicit <font face="Berkeley Book"> runs and per-paragraph <div style="margin-*; text-indent: …"> containers). Embedded performance-chart imagery shifted from being occasional to being routine.
documentFormatFiles[] entry whose type == "497K3B" and sequence == "1"; primary filenames are filer-chosen and not predictable, so resolve them through the metadata rather than guessing.<R> ... </R> redline markers (which surround text changed relative to a prior amendment), and handling of stray hex-encoded glyph artefacts (e.g., Ð/Ñ em-dash residues from PDF-style typesetting). Section boundaries are most reliably located by matching the canonical question headings rather than by HTML heading tags, which are essentially absent.entities) and multi-series records (seriesAndClassesContractsInformation length greater than one) are common and must be handled — a 497K3B accession does not map one-to-one to a single fund, CIK, or ticker. The series/class registry is the authoritative bridge from accession to marketed fund and to per-class tickers.effectivenessDate, seriesAndClassesContractsInformation, and stateOfIncorporation may be missing on individual records and should be treated as nullable.documentFormatFiles[].documentUrl. The same fallback applies to the complete-submission .txt referenced by linkToTxt.<R> redline markers inside an HTML body indicate text changed relative to the prior version of that Profile, but reconstructing the change history requires assembling successive accessions for the same series/class set.The filer of a Form 497K3B is always the registered open-end management investment company itself — the registrant on a Form N-1A registration statement under the Securities Act of 1933 and the Investment Company Act of 1940. In the EDGAR submission, the filer is the trust, corporation, or other legal vehicle (commonly a Massachusetts business trust, Delaware statutory trust, or Maryland corporation) that holds the registration; the individual fund series and share classes named in the Profile are the subjects of disclosure, not the filer of record.
The submission is signed and transmitted by an officer of the registrant or by a filing agent acting on its behalf. The investment adviser, sub-adviser, principal underwriter, distributor, and transfer agent are described inside the Profile but never file the 497K3B in their own capacity.
The reporting population is narrow. The form was used only by traditional open-end mutual funds. The following adjacent entity types did not file 497K3B and used different forms or sub-codes:
Profile use was always optional. A fund could comply with prospectus delivery obligations by delivering only its full statutory prospectus and never file a 497K3B; the absence of filings for a given fund family during 1998–2009 implies no compliance gap.
Form 497K3B sat at the intersection of two rules under the Securities Act of 1933:
The underlying registration statement to which a 497K3B always relates is Form N-1A. Profile content must conform to the corresponding N-1A disclosures.
A 497K3B filing is event-driven, not periodic. Rule 497 requires that the definitive form of any Profile actually used with investors be filed with the Commission, generally no later than the date of first use. Typical triggers are:
There is no fixed calendar deadline unique to 497K3B. Filings tend to cluster around each registrant's annual prospectus update cycle.
The dataset's temporal boundaries are set by the lifespan of the Profile regime:
Because the form is retired, the dataset is intrinsically closed: no new 497K3B filings are possible, and no pre-EDGAR paper records exist (the Profile regime never operated on paper).
Form 497K3B sits inside a dense cluster of mutual fund offering filings made under Rule 497 of the Securities Act, all anchored to the Form N-1A registration regime. Because 497K3B was a narrow, time-limited Profile format used from October 1998 through its discontinuation on March 30, 2009, the most informative comparisons are with its Rule 497 siblings, the N-1A registration it derives from, the 485-series amendments that drive its updates, the post-2009 summary prospectus regime that replaced it, and the periodic disclosure forms it is sometimes confused with.
Form 497 is the broad Rule 497 channel for definitive prospectuses, stickers, and supplements. A single 497 submission may carry a full statutory prospectus, a one-paragraph portfolio manager update, or a multi-series combined prospectus. 497K3B is a tightly scoped sub-form carrying only the standardized Profile defined by the 1998 Rule 498. Pulling 497 yields a heterogeneous, textually variable corpus; 497K3B yields a narrow, format-controlled subset.
The 497K-series sub-codes all carried fund Profiles built on the same 1998 Rule 498 schema (objectives, risks, fee table, performance, purchase/redemption mechanics). The trailing digits and letters distinguish the targeted distribution context (e.g., new-investor Profiles versus Profiles tied to specific share-class or acquirer scenarios), not the underlying disclosure architecture. Researchers studying 1998–2009 Profiles typically need all four sub-codes together. Note that the bare code "497K" was repurposed in April 2009 as the EDGAR submission type for the new summary prospectus, which is a legally different document despite the shared code.
497AD covers Rule 482 advertising materials. It shares the Rule 497 channel and a retail audience with 497K3B, but the content and legal status diverge sharply: 497AD is marketing communication subject to Rule 482's performance and disclosure rules, while 497K3B was a Section 10 offering document that satisfied prospectus delivery. A 497AD never could.
N-1A is the foundational registration for open-end funds, containing the full statutory prospectus, the SAI, and exhibits. 497K3B is a derivative summary of the corresponding N-1A, restricted to a defined subset of items. N-1A holds the legally definitive, long-form disclosure (including SAI-only items such as director compensation, control persons, and tax matters); the Profile never carried these. The two complement each other but cannot substitute.
The 485 series updates an effective N-1A registration, typically annually or on material change, and carries the refreshed prospectus, SAI, and financials. A new 485BPOS effective date generally triggered a refreshed Profile under one of the 497K codes. The 485 series is upstream and registration-oriented, tied to staff review and effectiveness; 497K3B is downstream, definitive, and delivery-oriented.
The summary prospectus regime that took effect March 31, 2009 replaced the 1998 Profile. The new Form 497K document looks superficially similar (key information, fees, risks, performance, management, purchase/sale) but operates under layered disclosure: the statutory prospectus must be posted online and is incorporated by reference. The 1998 Profile, by contrast, was a stand-alone Section 10 document that gave investors an explicit choice to buy on the Profile or request the full prospectus first. For longitudinal research, 497K3B covers 1998–2009 and post-2009 497K continues the series, but the required sections, delivery mechanics, and liability framework differ and the two should not be merged uncritically.
These share a filer population with 497K3B but cover a different disclosure axis. N-CSR carries certified shareholder reports with audited financials; N-Q (2004–2019) and N-PORT (its successor) carry portfolio holdings. None are offering documents and none contain a fee table or purchase/redemption procedures. They are useful complements for pairing offering-time disclosure with later holdings and performance, but they do not substitute for the Profile.
Form 497K3B is the dedicated archive of one specific Profile sub-code filed under the original 1998 Rule 498 from October 1998 through March 30, 2009. It is narrower than the full Rule 497 channel, distinct in audience and lifecycle from its 497K sibling sub-codes, derivative of N-1A and the 485 series, and legally separate from the post-2009 497K summary prospectus that shares its code. It has no overlap with the holdings-focused N-CSR, N-Q, or N-PORT regime. Any reconstruction of a fund's full disclosure history during this window requires pairing 497K3B with the corresponding N-1A, 485-series, and other 497K filings.
A closed archive of short-form mutual fund Profiles filed on EDGAR between October 1998 and March 2009, this dataset serves users working on historical research, regulatory archives, legacy litigation, NLP corpus building, and policy precedent rather than real-time investment workflows.
Counsel use the archive as primary-source precedent for interpreting today's Rule 498 summary prospectus regime. They mine the plain-English objective, principal-strategy, risk, and fee-table sections to trace how the Profile framework shaped current obligations, supporting interpretive memos, regulatory comment letters, and drafting guidance grounded in the historical rule.
Recordkeeping and compliance teams pull the metadata file and full document set in each accession to reconcile internal disclosure libraries with the public record. The workflow satisfies books-and-records obligations and supports responses to regulator inquiries reaching into the pre-2009 disclosure history of funds that still exist under merged or renamed share classes.
Teams working on arbitration, class-action, or enforcement matters tied to legacy fund products use the dataset to retrieve the exact Profile in effect on a given date. Filing-date metadata, risk-factor language, fee disclosures, and adviser descriptions become document-production exhibits and the basis for expert reports reconstructing what investors were told.
Researchers in financial economics and securities-disclosure history treat the 1998 to 2009 window as a natural experiment in which long-form prospectuses and short-form Profiles coexisted. They study fee tables, risk language, and performance presentations to test questions about flow effects, language convergence, and disclosure design, feeding peer-reviewed and working papers on regulatory reform.
Policy analysts at securities regulators, SROs, and policy research groups examine the Profile experiment to inform ongoing debates on layered disclosure, plain-English standards, and retail comprehension. Standardized fee tables and risk-section structures supply the empirical base for white papers and rule-review work on disclosure modernization.
Engineers building financial-document training corpora value the dataset as a bounded, structurally consistent collection. The HTML and TXT documents support labeled training sets for prospectus section classification, fee-table extraction, and risk-factor tagging, and the finite scope makes it a useful held-out evaluation corpus for models trained on modern summary prospectus text.
Marketing and product teams at asset managers retrieve the original objective wording, strategy descriptions, performance histories, and portfolio-manager bios for funds that have been renamed, merged, or repositioned. Outputs include brand-history projects, anniversary materials, and benchmarking of current summary prospectus voice against the Profile-era original.
Forensic accountants reviewing historical fund operations cross-reference the adviser, sub-adviser, fee, and principal-strategy disclosures against later filings to detect undisclosed strategy drift, fee changes, or personnel turnover material to an investigation.
Information-science professionals responsible for financial document archives integrate the dataset into broader EDGAR holdings as a complete, bounded corpus for a discontinued form, providing research access for downstream legal, academic, and journalistic users.
The Form 497K3B archive supports a small set of historical, archival, and corpus-building workflows that depend on its closed scope and tightly templated Profile structure.
Litigation-support teams working on arbitration or class-action matters tied to legacy fund purchases retrieve the exact Profile language an investor would have received on a specific transaction date. The workflow keys on filedAt and effectivenessDate in metadata.json, the seriesAndClassesContractsInformation[] registry to confirm the relevant series and ticker, and the <R>...</R> redline markers in the primary HTML to identify what changed from the prior amendment. The output is an authenticated production exhibit and a chronology of disclosure deltas for an expert report.
Investment-management attorneys advising on current Rule 498 summary prospectus obligations mine the eleven canonical question-form sections (objective, principal strategy, main risks, fees, management, purchase, redemption, distributions, services) for precedent phrasing. Because the Profile template is fixed across the entire 1998 to 2009 window, the corpus supports apples-to-apples comparison of how specific risk categories or fee-waiver constructions were originally drafted under the predecessor regime, feeding interpretive memos and SEC comment letters.
Corpus engineers use the dataset as a bounded, format-controlled training and held-out evaluation set for models that classify prospectus sections, extract Shareholder Fees and Annual Fund Operating Expenses tables, and tag risk-factor paragraphs. The fixed question headings provide reliable section labels, the standardized expense tables and Expense Example provide structured extraction targets, and the closed cutoff date prevents test-set leakage from models trained on modern post-2009 497K summary prospectuses.
Financial historians and policy researchers parse the Annual Fund Operating Expenses table, Average Annual Total Returns table, and risk enumeration across successive amendments for the same series (joined via seriesAndClassesContractsInformation[].series) to measure how management fees, 12b-1 charges, and risk-factor inventories evolved across the decade. The output supports working papers on the 2001 after-tax-return amendment, on risk-factor proliferation, and on the natural experiment of Profile-and-prospectus coexistence.
Compliance archivists at fund complexes pull each accession's documentFormatFiles[] inventory and primary 497K3B HTML to reconcile internal disclosure libraries against the EDGAR record for funds that still operate under merged, renamed, or reorganized share classes. The entities[] block (CIK, IRS number, file number, state of incorporation) and the series/class registry are the join keys back to current fund identifiers, supporting books-and-records obligations and regulator-inquiry responses reaching into the pre-2009 disclosure history.
Product and shareholder-communications teams retrieve the original objective sentence, strategy tagline, portfolio-manager biographies and tenure, and adviser/distributor block from the closing pages of the Profile to reconstruct the launch-era voice of funds that have since been rebranded or absorbed. The same fields support forensic reviews that cross-reference advisers, sub-advisers, and stated principal strategies against later 485BPOS amendments to flag undisclosed strategy drift or personnel turnover.
The Form 497K3B Files Dataset is accessible through three endpoints: a free metadata index, a full archive download, and per-container downloads.
Dataset Index JSON API: https://api.sec-api.io/datasets/form-497k3b-files.json
Returns dataset-level metadata (name, description, last update timestamp, earliest sample date, total records and size, form types, container format, and file types) along with the full list of container files. Each container entry includes its size, record count, last updated timestamp, and direct download URL. This endpoint does not require an API key and can be polled to detect which containers were refreshed in a given run, so downstream pipelines can fetch only the changed containers instead of re-downloading the entire dataset.
Example response:
1
{
2
"datasetId": "1f13365b-9ae0-697a-8739-d4331d24ac6d",
3
"datasetDownloadUrl": "https://api.sec-api.io/datasets/form-497k3b-files.zip",
4
"name": "Form 497K3B Files Dataset",
5
"updatedAt": "2026-04-15T08:05:52.174Z",
6
"earliestSampleDate": "1998-10-01",
7
"totalRecords": 2301,
8
"totalSize": 33999867,
9
"formTypes": ["497K3B"],
10
"containerFormat": "ZIP",
11
"fileTypes": ["TXT", "JSON", "HTML"],
12
"containers": [
13
{
14
"downloadUrl": "https://api.sec-api.io/datasets/form-497k3b-files/2009/2009-03.zip",
15
"key": "2009/2009-03.zip",
16
"size": 1248301,
17
"records": 87,
18
"updatedAt": "2026-04-15T08:05:52.174Z"
19
}
20
]
21
}
Download Entire Dataset: https://api.sec-api.io/datasets/form-497k3b-files.zip?token=YOUR_API_KEY
Downloads the complete dataset as a single ZIP archive covering all Form 497K3B filings from October 1998 through the form's discontinuation in March 2009. This endpoint requires an API key.
Download Single Container: https://api.sec-api.io/datasets/form-497k3b-files/2009/2009-03.zip?token=YOUR_API_KEY
Downloads one monthly container ZIP rather than the full archive. Use the downloadUrl values from the index JSON to retrieve specific months. This endpoint requires an API key.
The dataset covers Form 497K3B, the EDGAR submission type used to file a mutual-fund "Profile" prepared in accordance with former Rule 498 — specifically, profiles described in former Rule 497(k)(1)(iii)(B). It is the dedicated archive of this one Profile sub-code and does not include sibling sub-codes such as 497K1, 497K2, or 497K3A.
A single record corresponds to one accepted EDGAR submission of Form 497K3B by a registered open-end management investment company. Each record is a folder named with the 18-digit accession number containing a metadata.json header file and the primary 497K3B HTML document; a single record may cover multiple registered series and share classes when a fund-family Profile bundles related funds into one accession.
The filer is always the registered open-end management investment company itself — the trust, corporation, or other legal vehicle that holds the Form N-1A registration. Profile use was always optional under former Rule 498, so filings reflect only those fund families that chose to deliver a Profile rather than relying solely on the full statutory prospectus.
The dataset spans October 1998, when the earliest 497K3B filings appeared on EDGAR following the SEC's adoption of Rule 498 in Release No. 33-7513, through March 30, 2009, when Release No. 33-8998 discontinued the original Profile regime. Because the form is retired, the dataset is intrinsically closed and no new 497K3B filings are possible.
The post-2009 Form 497K implements the summary prospectus framework under amended Rule 498, which uses layered disclosure (the statutory prospectus is posted online and incorporated by reference). The 1998 Profile captured here was a stand-alone Section 10 offering document that gave investors an explicit choice to buy on the Profile or first request the full prospectus. The two forms share an EDGAR code lineage but operate under different statutory mechanics and should not be merged uncritically for longitudinal research.
The dataset is distributed as monthly ZIP containers organized by year (paths of the form <YYYY>/<YYYY-MM>.zip). Inside each container, records use TXT, JSON, and HTML file types — a flat metadata.json header per accession, the primary 497K3B HTML document, and any short ancillary text documents that accompanied the original submission. GIF images referenced from the HTML and the complete-submission .txt SGML bundle are intentionally excluded; URLs in the metadata point back to the original EDGAR archive copies.
Amendments produce new 497K3B accessions rather than in-place edits to an existing record. Inside the HTML body of an amendment, literal <R> ... </R> redline markers surround text changed relative to the prior version of the Profile, but reconstructing the full change history requires assembling successive accessions for the same series/class set using the seriesAndClassesContractsInformation[] registry as the join key.