Form N-CSRS Files Dataset

The Form N-CSRS Files dataset is a collection of every certified semi-annual shareholder report filed on Form N-CSRS or Form N-CSRS/A by registered management investment companies — open-end mutual funds, closed-end funds, and ETFs organized as registered investment companies — under Section 30(b)(2) of the Investment Company Act of 1940 and Rule 30e-1 thereunder. Each record is a single EDGAR accession number packaged as an accession-numbered folder containing the primary semi-annual report, the Sarbanes-Oxley Section 302 and Section 906 officer certifications, any optional EX-99 attachments, and a metadata.json spine that mirrors the EDGAR submission header. The dataset begins on 2003-07-01 — the introduction of N-CSRS in the Commission's Sarbanes-Oxley implementation for investment companies — and is refreshed continuously as new accessions arrive on EDGAR. Files are distributed in TXT, JSON, HTML, and PDF inside monthly ZIP containers organized as <year>/<year>-<month>.zip. The dataset captures both the legacy SGML-wrapped HTML era and the modern inline-XBRL era introduced by the SEC's 2019/2020 SEC tagging rules and the 2022 Tailored Shareholder Reports rule.

Dataset APIs

Dataset Files

What This Dataset Contains

The dataset packages the population of Form N-CSRS and Form N-CSRS/A filings on EDGAR from July 2003 forward into monthly ZIP containers. Form N-CSRS is the certified semi-annual shareholder report mandated by Section 30(b)(2) of the Investment Company Act of 1940 and Rule 30e-1 thereunder, filed by registered management investment companies within ten days after transmitting the semi-annual report to shareholders. The filing's central content is the semi-annual shareholder report itself — the schedule of investments together with the unaudited financial statements for the half-year period ending on the registrant's fiscal mid-point — bundled with the principal-officer certifications introduced by the Sarbanes-Oxley Act of 2002 (Sections 302 and 906) and the standard EDGAR mechanics (form cover page, signatures, exhibit index).

Form N-CSRS shares its item structure with Form N-CSR (the annual variant). The semi-annual variant differs in that several N-CSR items are not required at the half-year point (notably the Code of Ethics, audit committee financial expert, principal accountant fees, and the auditor-related controls disclosures), the financial statements are unaudited, and no auditor's report is included. A Form N-CSRS/A record is an amendment to a previously filed N-CSRS — structurally indistinguishable from a Form N-CSRS record except that the formType value in metadata.json reads "N-CSRS/A" and the body conventionally restates or supersedes the prior submission.

Files are distributed as TXT, JSON, HTML, and PDF inside ZIP containers organized as <year>/<year>-<month>.zip. The dataset preserves everything EDGAR received in each submission except for image attachments and standalone XBRL taxonomy/linkbase sidecars; the inventory of those omitted files remains traceable through metadata.json.

Content Structure of a Single Record

1. What one record represents

One record in the Form N-CSRS Files dataset is a single EDGAR accession of Form N-CSRS or Form N-CSRS/A — that is, one certified shareholder report (or one amendment to such a report) filed by a registered management investment company. Each record materializes on disk as a single accession-numbered folder whose name is the bare 18-digit accession number with no separators (for example 000119312525260531). Inside the folder live exactly one metadata.json, the primary report document, the Sarbanes-Oxley certification exhibits, and any optional EX-99 attachments that the registrant submitted. Images attached to the original submission are excluded from the folder, and the standalone XBRL taxonomy/linkbase sidecars are excluded as well; everything else EDGAR received in the submission is present.

A single record may cover one fund, several series of a multi-series investment company filing one report at once, or — for N-CSRS/A records — a corrected version of a previously transmitted semi-annual report. The accession folder is the canonical unit of analysis. Container partitioning into yearly and year-month ZIPs (<year>/<year>-<month>.zip) is purely a packaging convenience and does not affect what constitutes one record.

2. Container and folder layout

The dataset is delivered as ZIP containers organized by calendar year and within each year by year-month. A record's location resolves to <year>/<year>-<month>.zip, and inside the ZIP the records sit as flat sibling folders directly under a <year>-<month>/ root. There is no per-CIK, per-series, or per-filer grouping inside the container, and there are no nested directories inside an accession folder. The result is a strict two-level hierarchy: monthly partition, then accession folder, then files.

The file types found in the dataset are TXT, JSON, HTML, and PDF. In practice, modern filings are dominated by HTML primary reports and HTML exhibit certifications plus the JSON metadata file; legacy filings may carry a .txt primary or PDF supplements, but the structural layout is identical regardless of body format. Filenames inside an accession folder are not standardized — each filing-agent template (Donnelley/Edgar Online, DFIN/ArcFiling, in-house systems) applies its own naming convention. The role of each file is always recoverable through the type discriminator inside metadata.json.documentFormatFiles[*], so consumers should rely on metadata typing rather than filename pattern-matching.

3. The metadata.json spine

metadata.json is the EDGAR submission header reshaped as JSON. Exactly one such file exists per accession folder, and it serves as the record's spine: every other file in the folder corresponds to one entry in documentFormatFiles (or to one entry in dataFiles, in the case of XBRL sidecars that are referenced but not extracted). The meaningful top-level fields include:

• formType — the literal "N-CSRS" or "N-CSRS/A".
• accessionNo — the dashed 18-digit EDGAR accession (e.g. 0001193125-25-260531), which de-dashed equals the folder name.
• filedAt — the ISO datetime with timezone offset of EDGAR acceptance.
• effectivenessDate — the date the filing became effective (typically the same calendar day as filedAt).
• periodOfReport — the end date of the semi-annual reporting period (e.g. 2025-08-31).
• description — the human label "Form N-CSRS - Certified Shareholder Report, Semi-Annual".
• linkToFilingDetails — the canonical EDGAR URL of the primary document (uses the /ix?doc=... viewer for inline-XBRL filings).
• linkToHtml — the EDGAR *-index.htm filing-index page.
• linkToTxt — the EDGAR <accession>.txt complete-submission text file.
• linkToXbrl — empty for N-CSRS, because the XBRL payload is embedded inline in the primary HTML rather than provided as a standalone instance.
• id — an internal 32-hex record identifier.
• documentFormatFiles — an array describing every primary, exhibit, and graphic document that EDGAR received in the submission.
• dataFiles — an array describing the XBRL sidecar files (schema, label, definition, presentation, and instance) referenced by the inline-XBRL primary document.
• entities — an array of one or more filer-entity blocks parsed from the EDGAR header.
• seriesAndClassesContractsInformation — an array enumerating the SEC Series IDs and per-class Contract IDs covered by the filing.

documentFormatFiles[*]

Each element describes one item EDGAR received and carries sequence (EDGAR's submission-order index, with "1" always the primary report), size (stringified byte count), documentUrl (the canonical EDGAR archive URL whose basename matches the on-disk filename for non-graphic types), description (a free-form label, often the registrant name or an exhibit caption), and type (the role discriminator). The type values that classify N-CSRS files are:

• "N-CSRS" — the primary report document.
• "EX-99.CERT" — the Sarbanes-Oxley Section 302 certification.
• "EX-99.906 CERT" (occasionally serialized without the space as "EX-99.906CERT") — the Section 906 certification.
• "EX-99" — any other exhibit (auditor letters, additional schedules, board-approval statements).
• "GRAPHIC" — embedded image attachments (referenced by the HTML but stripped from the ZIP).

The complete-submission text file appears at the end of the array with a blank sequence and the description "Complete submission text file"; it is referenced by URL but the file itself is not present in the folder.

dataFiles[*]

This array uses the same shape as documentFormatFiles but the type values are XBRL-specific: "EX-101.SCH" (taxonomy schema), "EX-101.DEF" (definition linkbase), "EX-101.LAB" (label linkbase), "EX-101.PRE" (presentation linkbase), and "XML" (extracted instance document). These entries exist for traceability — they document what EDGAR received — but the underlying files are not extracted into the accession folder, because the inline-XBRL facts are already serialized inside the primary HTML.

entities[*]

One object per filer or co-filer participating in the submission. Common keys are companyName (with a trailing role tag such as " (Filer)"), cik, fileNo (the Investment Company Act file number, e.g. "811-03712"), irsNo, stateOfIncorporation (two-letter), act (always "40" for the Investment Company Act of 1940), fiscalYearEnd (MMDD format, e.g. "0228"), type (the form type for this filer's role in the submission), and filmNo (the EDGAR film number).

seriesAndClassesContractsInformation[*]

For management investment companies, this array enumerates the SEC Series IDs and per-class Contract IDs covered by the filing. Each element carries the series identifier (series, e.g. "S000004573"), the series name, and a classesContracts sub-array of { ticker, name, classContract } triples — one per share class. A single N-CSRS filing can cover multiple series and arbitrarily many share classes; these Class IDs are the same identifiers that the inline-XBRL contexts inside the primary HTML reference via the oef:ClassAxis dimension.

4. The primary N-CSRS report document

The primary document (type: "N-CSRS" in documentFormatFiles, always sequence "1") is the certified shareholder report itself, rendered as an HTML file. For modern filings it is also valid inline XBRL (iXBRL): the same file simultaneously serves as the human-readable report and as the machine-readable XBRL instance, with the structured facts woven into the HTML through <ix:*> elements.

Cover page and EDGAR scaffolding

The body opens with the standard EDGAR cover scaffolding for N-CSRS: the "United States Securities and Exchange Commission, Washington, D.C. 20549, FORM N-CSR, Certified Shareholder Report of Registered Management Investment Companies" caption, the file-number block (Investment Company Act file number 811-XXXXX), the registrant's exact legal name and principal-office address, the name and address of the agent for service of process, the registrant's telephone number, fiscal year end, and the date of the reporting period covered. A check-box block flags the filing as a semi-annual report.

Inline-XBRL header block

A modern N-CSRS report opens with an XHTML root that imports the inline-XBRL namespaces (ix, xbrli, link, xlink) and the relevant SEC and FASB taxonomies — dei (Document and Entity Information), oef (Open-End Fund Reporting), and us-gaap for the dated taxonomy versions. Immediately inside the body, a hidden <div> holds an <ix:header> block containing three sub-blocks:

• <ix:hidden> — non-numeric facts that anchor the document, such as document type (N-CSRS), the registrant's CIK, the period flag (semi-annual shareholder report), and similar dei/oef tags.
• <ix:references> — a <link:schemaRef> pointing to the schema sidecar file (e.g. cik0000717819-20250831.xsd).
• <ix:resources> — the <xbrli:context> and <xbrli:unit> definitions used throughout the body.

Contexts are factored along several axes — most prominently oef:ClassAxis (one context per share-class Contract ID) and the various concentration axes (oef:CreditQualityAxis, oef:GeographicConcentrationAxis, oef:IndustrySectorAxis, etc.). A single filing covering several share classes can therefore produce hundreds of context blocks, each with its own period and dimensional segment. Numeric facts in the visible body are tagged with <ix:nonFraction> and reference these contexts and units; non-numeric facts in the visible body use <ix:nonNumeric>.

Item-by-item body

After the cover scaffolding and the inline-XBRL header, the document presents the standard Form N-CSR item structure rendered as styled <div> and <table> blocks. At semi-annual reporting, the substantive content concentrates in Items 1 and 7; the remaining items are either cross-references, "Not applicable" placeholders carried for structural consistency, or closed-end-fund-specific blocks that only certain registrants populate.

• Item 1 — Reports to Stockholders. The shareholder report transmitted under Rule 30e-1, attached or restated in full. For open-end funds subject to the Tailored Shareholder Reports rule this is the streamlined per-class TSR document; for closed-end funds and other registrants it is the traditional narrative report (portfolio manager commentary, performance discussion, expense disclosures, fund-overview tables, and the schedule of investments).
• Item 2 — Code of Ethics. Generally not required at semi-annual; appears as a "Not applicable" placeholder.
• Item 3 — Audit Committee Financial Expert. Generally not required at semi-annual.
• Item 4 — Principal Accountant Fees and Services. Generally not required at semi-annual.
• Item 5 — Audit Committee of Listed Registrants. Generally not required at semi-annual; applicable only to listed closed-end funds at annual.
• Item 6 — Investments. (a) The schedule of investments included as part of the report to stockholders under Item 1. (b) Material changes in investments since the last report, where applicable.
• Item 7 — Financial Statements and Financial Highlights for Open-End Management Investment Companies. Following the 2022 Tailored Shareholder Reports rule, the unaudited financial statements and financial highlights for open-end funds are presented under this item rather than inside the shareholder-report letter. The block includes the Statement of Assets and Liabilities, the Statement of Operations, the Statement of Changes in Net Assets, the per-share financial highlights tables for each share class, the detailed schedule of investments (security-by-security holdings with CUSIPs, par/share quantities, fair values, and percent-of-net-assets allocations), and the notes to financial statements (accounting policies, fair-value hierarchy, derivative disclosures, related-party transactions, federal tax information, subsequent events).
• Items 8–10 — Open-end fund operating disclosures. Changes in and disagreements with accountants, proxy disclosures of fund expenses paid in connection with the report, statement regarding the basis for the board's approval of investment advisory contracts (typically presented at the meeting at which renewal occurs).
• Items 11–14 — Closed-end fund disclosures. Statement regarding basis for approval of investment advisory contract, financial statements and financial highlights for closed-end management investment companies (where presented separately), purchases of equity securities by the closed-end fund and affiliated purchasers, portfolio managers of closed-end management investment companies, and submission of matters to a vote of security holders.
• Item 15 — Submission of Matters to a Vote of Security Holders. Disclosure of any material changes to procedures by which shareholders may recommend nominees to the board.
• Item 16 — Controls and Procedures. The principal officers' conclusions about the effectiveness of the registrant's disclosure controls and procedures and any changes in internal control over financial reporting that materially affected, or are reasonably likely to materially affect, the registrant's internal control over financial reporting during the period.
• Item 17 — Disclosure of Securities Lending Activities for Closed-End Management Investment Companies.
• Item 18 — Recovery of Erroneously Awarded Compensation. Disclosure of compensation-recovery actions under the Dodd-Frank clawback rules, where applicable.
• Item 19 — Exhibits. The exhibit index that references the certifications and any other EX-99 attachments physically delivered as separate files in the submission.

Item presence and exact ordering vary by registrant type (open-end vs. closed-end vs. ETF), by filing era, and by whether the registrant is subject to the Tailored Shareholder Reports rule. The set of oef: tagged facts present in the inline-XBRL header similarly varies with which TSR disclosures apply.

Signatures

The primary document closes with a SIGNATURES block: the registrant's exact legal name, the date, and /s/ <name> signature lines for the principal executive officer and the principal financial officer (separately from the certification exhibits). The signatures block is part of the primary HTML body and is not stored as a separate file.

Legacy SGML-wrapped HTML

Some filings, particularly from older periods or smaller filing agents, deliver the primary document as HTML wrapped in EDGAR's SGML <DOCUMENT> envelope rather than as a clean XHTML root. In those cases the body opens with <TYPE>, <SEQUENCE>, <FILENAME>, <DESCRIPTION> headers followed by <TEXT>...</TEXT> containing the actual HTML. This wrapper is identical in form to the wrapper used by the certification exhibits described next.

5. The Sarbanes-Oxley certification exhibits

Two certification exhibits are present in essentially every N-CSRS filing: the Sarbanes-Oxley Section 302 certification (EX-99.CERT) and the Sarbanes-Oxley Section 906 certification (EX-99.906 CERT). Both are short HTML documents wrapped in EDGAR's SGML <DOCUMENT> header.

EX-99.CERT — Section 302

The Section 302 certification is the principal-executive-officer and principal-financial-officer attestation mandated by Section 302 of the Sarbanes-Oxley Act, as implemented for registered investment companies under Rule 30a-2(a) and Rule 30a-3 under the Investment Company Act. The body of the exhibit recites the standard five-paragraph attestation: that the certifying officer has reviewed the report; that based on their knowledge the report contains no untrue statement of material fact and omits no material fact necessary to make the statements not misleading; that the financial statements and other financial information fairly present in all material respects the financial condition, results of operations, and changes in net assets; that the certifying officers are responsible for establishing and maintaining disclosure controls and procedures (Rule 30a-3(c)) and internal control over financial reporting (Rule 30a-3(d)) and have designed, evaluated, and disclosed any material weaknesses in those controls; and that any fraud involving management or employees with a significant role in internal control has been disclosed. Two parallel certifications appear when the principal executive officer and the principal financial officer sign separately.

EX-99.906CERT — Section 906

The Section 906 certification is the separate attestation required under 18 U.S.C. § 1350 (Sarbanes-Oxley Section 906). It is shorter than the Section 302 exhibit and recites that the periodic report fully complies with the requirements of Section 13(a) or 15(d) of the Securities Exchange Act of 1934 (or, for investment-company N-CSRS filings, the analogous provisions) and that the information contained in the report fairly presents in all material respects the financial condition and results of operations of the registrant. The certification is signed by both the principal executive officer and the principal financial officer, typically with /s/ <name> signature lines. Unlike the Section 302 exhibit, the Section 906 certification is "furnished" rather than "filed."

SGML wrapper

Both certifications follow the same SGML wrapper pattern: a <DOCUMENT> header with <TYPE>EX-99.CERT (or EX-99.906 CERT), <SEQUENCE>, <FILENAME>, and <DESCRIPTION> lines, followed by <TEXT> containing the HTML body and a closing </TEXT></DOCUMENT>.

6. Optional auxiliary EX-99 exhibits

A subset of filings carry one or more additional EX-99 exhibits beyond the two mandatory certifications. Common examples include independent-auditor letters to the SEC commenting on Item 8 (changes in or disagreements with accountants), supplementary schedules referenced from the body, statements regarding the board's approval of investment advisory contracts, and — for certain closed-end and listed funds — proxy results or notices of shareholder meetings. These auxiliary attachments use the same SGML <DOCUMENT> wrapper as the certifications, with <TYPE>EX-99 and a <DESCRIPTION> matching the exhibit's purpose. They appear with sequence values after the primary report and certifications in documentFormatFiles.

7. What the dataset record includes vs. omits

A record includes:

1. metadata.json — the EDGAR submission header reshaped as JSON.
2. The primary N-CSRS report document (HTML in modern filings; TXT in some legacy filings).
3. The EX-99.CERT Section 302 certification.
4. The EX-99.906 CERT Section 906 certification.
5. Any auxiliary EX-99 exhibits originally submitted (e.g. auditor letters, advisory-contract approval statements).

A record omits two categories of files that exist in the original EDGAR submission and remain listed in metadata.json:

1. Image attachments (entries with type: "GRAPHIC" in documentFormatFiles, typically JPG/PNG/GIF used for fund logos, performance charts, allocation pie charts, and similar). These are referenced by <img src="..."> tags in the primary HTML, but the binary image files are stripped. The HTML therefore renders with broken image references.
2. XBRL data sidecars (entries in dataFiles with types EX-101.SCH, EX-101.DEF, EX-101.LAB, EX-101.PRE, and XML). The taxonomy and linkbase files referenced by the inline-XBRL header are listed in metadata.json but are not extracted into the folder, since the inline-XBRL facts themselves live inside the primary HTML.

The complete-submission .txt (EDGAR's SGML concatenation of every document in the submission) is also referenced by linkToTxt and listed at the tail of documentFormatFiles but is not stored in the folder.

The full inventory of every original document — kept or stripped — is preserved in metadata.json, so consumers can reconstruct the complete EDGAR submission inventory and pull omitted files directly from the EDGAR archive URLs in documentFormatFiles[*].documentUrl and dataFiles[*].documentUrl.

8. Evolution of structure and format over time

Form N-CSRS was adopted in 2003 as part of the SEC's implementation of the Sarbanes-Oxley Act for registered investment companies, and the dataset begins with filings from July 2003. Several material structural and format shifts have occurred since:

• Early SGML era (2003 through the mid-2000s). The earliest filings were submitted as EDGAR SGML packages with the body and exhibits often delivered as plain-text or minimally-formatted HTML inside <DOCUMENT> SGML wrappers. The Section 302 and 906 certifications were already mandatory from the form's inception.
• HTML adoption and richer formatting (mid-2000s through the 2010s). Filings transitioned to richer HTML primary documents with inline CSS and embedded image graphics, while the certification exhibits continued to use the SGML-wrapped HTML pattern still seen today.
• Item additions tied to subsequent rulemaking. Items added or amended over time include disclosures on the audit committee's pre-approval policies, refinements to the statement-of-changes presentation, securities-lending activities for closed-end funds (Item 17), and the Item 18 disclosure on recovery of erroneously awarded compensation pursuant to the Dodd-Frank clawback rules implemented through SEC and exchange listing standards.
• Inline-XBRL adoption following the 2019/2020 SEC tagging rules. The SEC's modernization rulemakings extended structured-data tagging to registered investment companies, and the 2022 Tailored Shareholder Reports rule (compliance dates in 2024 for most filers) required N-CSRS reports to be tagged with oef (Open-End Fund Reporting) facts inline. Modern filings therefore carry the <ix:header>/<ix:hidden>/<ix:resources> block at the top of the primary HTML, dimensional contexts factored along oef:ClassAxis and the various concentration axes, and tagged numeric facts throughout the financial statements and schedule-of-investments tables. The XBRL sidecar files (schema and linkbases) are still produced and listed in dataFiles, but the inline serialization in the primary HTML is what carries the actual fact data.
• Filename and packaging conventions. Filename stems vary by filing-agent template across the entire history of the form: Donnelley/Edgar Online use patterns like d<digits>dncsrs.htm, d<digits>dex99cert.htm, d<digits>dex99906cert.htm; DFIN/ArcFiling use patterns like <ticker>-<job>_ncsrs.htm, <ticker>-<job>_ex99cert.htm, <ticker>-<job>_ex99906cert.htm; in-house templates use short literal names such as output.htm, cert302.htm, cert906.htm; inline-XBRL primary reports are often suffixed _ncsrsixbrl.htm. Filenames should never be relied on as semantic discriminators — the documentFormatFiles[*].type field is the authoritative role indicator.

9. Interpretation and extraction notes

• Folder names are accession numbers, not filer identifiers. Multiple folders for the same filer in the same month are not grouped; consumers must join on entities[*].cik from metadata.json to aggregate by registrant.
• Multi-series filings. A single N-CSRS submission can cover multiple SEC Series IDs and many share-class Contract IDs. The seriesAndClassesContractsInformation array enumerates them, and the inline-XBRL contexts in the primary HTML are dimensioned along oef:ClassAxis using the same Class IDs. Per-class facts must be extracted by matching contextRef to the class dimension rather than assumed to be filing-wide.
• Amendments. Records with formType "N-CSRS/A" are structurally identical to their "N-CSRS" counterparts; the dataset does not link an amendment to the underlying filing it amends. Consumers needing the supersession relationship must reconstruct it from accessionNo, entities[*].cik, and periodOfReport.
• Broken image references in the primary HTML. Because GRAPHIC files are stripped from the ZIP, the primary HTML's <img> tags point to filenames that do not exist alongside them. Rendering the HTML in a browser shows missing images; the underlying binaries are obtainable from EDGAR via the URLs preserved in metadata.json.
• Inline XBRL vs. the empty linkToXbrl. linkToXbrl is empty for N-CSRS filings because the XBRL facts are not packaged as a standalone instance — they are inline within the primary HTML. Extraction therefore requires an iXBRL-aware parser operating directly on the primary document.
• The complete-submission .txt. EDGAR's combined submission text file is referenced by linkToTxt and listed at the tail of documentFormatFiles with a blank sequence, but the file itself is not present in the accession folder; consumers needing the SGML-wrapped concatenation must fetch it from EDGAR.
• Exhibit signatures. The Section 302 and 906 certifications carry the principal executive officer's and principal financial officer's signatures as /s/ <name> lines in the HTML body. There is no separate signature file; signatures are part of each exhibit's body. The primary report carries its own separate SIGNATURES block.
• Form-type spelling for Section 906 exhibits. The type value is sometimes serialized as "EX-99.906 CERT" (with a space) and sometimes as "EX-99.906CERT" (without). Consumers filtering by exhibit type should normalize whitespace.
• SGML envelope handling. The certification exhibits — and some legacy primary documents — are HTML wrapped in EDGAR's SGML <DOCUMENT> header. HTML parsers fed these files raw will see the SGML preamble as text; consumers should either strip the wrapper to the inner <TEXT>...</TEXT> body or use a parser tolerant of the envelope.

Who Files or Publishes This Dataset, and When

Who files

Form N-CSRS is filed by registered management investment companies (as defined in Sections 4 and 5 of the Investment Company Act of 1940). The legal filer is the registrant itself — the trust, corporation, or partnership — acting through its principal executive officer and principal financial officer, who sign the Sarbanes-Oxley certifications. The investment adviser, distributor, transfer agent, custodian, and administrator are not the filing parties, even though they typically generate the underlying data. The board and audit committee oversee the financial statements but do not bear the filing obligation.

The reporting population includes:

• Open-end management investment companies (mutual funds), including series trusts where each series has its own fiscal calendar.
• Closed-end management investment companies, including listed closed-end funds, interval funds, and tender-offer funds.
• ETFs organized as open-end (or, less commonly, closed-end) management investment companies.

The unit of observation is the EDGAR accession number, not the individual series. A single N-CSRS filing by a series trust may cover multiple series whose semi-annual periods coincide.

Entities outside the N-CSRS population

The following do not file N-CSRS:

• Unit investment trusts (UITs) — not management companies; they file Form N-CEN and, where applicable, Form N-PX.
• Business development companies (BDCs) — although organized as closed-end funds, BDCs that have elected BDC status under Section 54 of the 1940 Act report under the Exchange Act regime (10-K, 10-Q, 8-K) and do not file N-CSRS.
• Face-amount certificate companies — separate 1940 Act category.
• Foreign funds not registered under the 1940 Act.
• Private funds relying on Section 3(c)(1) or Section 3(c)(7) exclusions.
• Variable insurance product separate accounts organized as UITs (separate accounts organized as management companies do file N-CSRS).
• SBICs that are not registered under the 1940 Act.

Statutory and regulatory framework

The obligation rests on a layered framework:

• Section 30(b)(2) of the Investment Company Act of 1940 — authorizes the SEC to require periodic reports from registered investment companies.
• Rule 30e-1 — requires every registered management investment company to transmit a semi-annual report containing specified financial information to each stockholder of record.
• Rule 30b2-1 — requires that every report transmitted under Rule 30e-1 be filed with the Commission on Form N-CSR (annual) or Form N-CSRS (semi-annual) within 10 days after transmission to stockholders.
• Rule 30a-2(a) — adopted under Sections 302 and 906 of the Sarbanes-Oxley Act of 2002, requires the principal executive officer(s) and principal financial officer(s) to provide the certifications attached to each N-CSRS.
• Rule 30a-3 — requires the registrant to maintain disclosure controls and procedures and internal control over financial reporting, addressed in the certifications.

Form N-CSRS was adopted in 2003 alongside the Commission's Sarbanes-Oxley implementation rules for investment companies, replacing Form N-30D for semi-annual shareholder reports.

Trigger and timing

N-CSRS is event-anchored to the transmission of the semi-annual shareholder report, not to a fixed calendar date:

1. The fund reaches the midpoint of its fiscal year and prepares the Rule 30e-1 semi-annual report.
2. The fund transmits the report to stockholders (by mail, or under Rule 30e-3 / tailored shareholder report rules, by website availability with notice).
3. Within 10 days after the transmission date, the registrant files Form N-CSRS on EDGAR, attaching the report as transmitted plus items required only in the EDGAR filing (Sarbanes-Oxley certifications, controls disclosures, and portfolio holdings information).

The 10-day clock runs from transmission, not from the close of the semi-annual period or the date the report is finalized internally. Because each registrant (and each series) may have a different fiscal year, N-CSRS filings appear continuously throughout the calendar year. A December 31 fiscal year-end produces a June 30 semi-annual period and an N-CSRS filing typically in late August or early September; an October 31 fiscal year-end produces an April 30 semi-annual period and an N-CSRS filing typically in July.

N-CSRS/A amendments

Form N-CSRS/A is an amendment to a previously filed Form N-CSRS, submitted by the same registrant. There is no fixed deadline; it is filed when the need is identified. Each N-CSRS/A receives a separate accession number and appears as a distinct record rather than as a replacement of the original. Typical reasons:

• correcting typographical, tagging, or numerical errors
• restating financial statements
• attaching a corrected or re-executed Sarbanes-Oxley certification
• supplying omitted exhibits or documents
• conforming the EDGAR document to a corrected version transmitted to stockholders

Important distinctions

• Legal filer vs. signatories — the registrant is the filer; the PEO and PFO are the certifying signatories.
• Series trusts — one N-CSRS may cover multiple series with aligned semi-annual periods; series with different fiscal year-ends produce multiple N-CSRS filings per trust per year.
• Money market funds file N-CSRS like other open-end funds; their separate Form N-MFP regime is unrelated to the semi-annual cycle.
• New, deregistered, or merged funds — N-CSRS begins after the first completed semi-annual period as a registered fund; a final N-CSRS may be owed for a partial period through deregistration or merger if a final shareholder report is transmitted.
• Tailored shareholder report (TSR) reforms changed the content and format transmitted to stockholders but did not alter the N-CSRS filing obligation or the 10-day trigger.

How This Dataset Differs From Similar Datasets or Filings

Form N-CSRS sits inside a tightly clustered family of Investment Company Act periodic reports. The most useful comparisons are the other shareholder-report and portfolio-holdings forms filed by registered management investment companies, plus the predecessor and successor regimes that bracket N-CSRS in time. Operating-company forms (Form 10-K, Form 10-Q) are not relevant comparators: different statute, filer population, and content.

Form N-CSR (annual certified shareholder report)

N-CSR is the direct annual analog and the closest comparator. Both are filed under Section 30(b)(2) and Rule 30e-1, both transmit the shareholder report to the SEC within ten days of mailing, both carry SOX 302 and 906 certifications, and both share the same Item shell.

Differences:

• Period and audit status. N-CSR covers the full fiscal year with audited financials and an auditor's report; N-CSRS covers the first six months with unaudited interim financials.
• Item coverage. Several Items attach only to N-CSR (notably principal accountant fees, certain code-of-ethics filings), so an N-CSRS accession is structurally lighter even when the schedule of investments is comparably detailed.
• Reconstruction. A complete annual disclosure cycle requires both; only N-CSRS captures the mid-year snapshot.

Form N-PORT (monthly portfolio holdings)

Form N-PORT (Rule 30b1-9) is the closest holdings-data substitute and the reason N-CSRS is no longer the primary source for portfolio composition. N-PORT is monthly, structured XML, with the third-month report of each fiscal quarter publicly disclosed after a 60-day lag. It carries position-level fields N-CSRS does not standardize: ISIN/CUSIP/LEI, fair value, percent of net assets, payoff profile, asset and issuer category, country, currency, derivatives reference data, liquidity classification, and risk metrics for certain funds.

What N-PORT does not carry: management discussion of performance, statement of assets and liabilities, statement of operations, statement of changes in net assets, financial highlights, notes (accounting policies, related-party transactions, tax, capital share activity, subsequent events), and SOX 302/906 certifications. N-PORT is a holdings tape; N-CSRS is a certified shareholder report. They are complements: N-PORT for granular holdings, N-CSRS for fund-level financial reporting and certified narrative.

Form N-Q (rescinded quarterly holdings)

Form N-Q filled the Q1/Q3 gap between N-CSRS and N-CSR from 2004 until rescission in 2018 when N-PORT public reporting took effect. Overlap with N-CSRS was confined to the schedule of investments; N-Q carried no financial statements, no notes, and no MD&A, and only narrower SOX certifications. Relevant today only for longitudinal studies covering 2004-2018.

Form N-30D (pre-2003 predecessor)

N-30D was the pre-Sarbanes-Oxley shareholder-report form that the 2003 reforms split into N-CSR and N-CSRS. Its embedded shareholder report was similar in scope, but it lacked the certified Item shell, the SOX 302/906 officer certifications, and the explicit code-of-ethics, audit-committee-financial-expert, and principal-accountant-fees disclosures. N-30D is the only backward continuation across the July 2003 cutover, which is exactly the N-CSRS dataset's earliest sample date (2003-07-01).

Form N-MFP (money market fund monthly portfolio)

N-MFP (Rule 30b1-7) applies only to money market funds and reports security-level holdings, WAM, WAL, daily/weekly liquid asset percentages, shadow NAV, and shareholder flows. It is high-frequency, narrow-population, supervisory, and entirely tabular. MMFs still file N-CSRS, which carries the narrative shareholder report, financial statements, and certifications that N-MFP omits. Complements, not substitutes: N-MFP for portfolio dynamics, N-CSRS for governance and accounting disclosure.

Tailored Shareholder Report regime (Rule 30e-3 and the 2022 reforms)

The 2018 Rule 30e-3 amendments and the 2022 TSR reforms changed delivery of shareholder content but did not eliminate the N-CSRS filing obligation. Under the 2022 reforms, open-end funds transmit a concise, fund- and class-specific TSR to shareholders, while full financial statements, the complete schedule of investments, and notes are posted online and filed with the SEC as exhibits to N-CSR/N-CSRS. Closed-end funds and certain other registrants remain outside TSR and continue to file the older N-CSRS structure.

Practical consequence for this dataset: post-TSR accessions split the historical N-CSRS payload across a short primary TSR document and one or more exhibits. The dataset captures both regimes in full, but parsers expecting a single monolithic primary document will mis-handle post-TSR filings.

Form N-PX (annual proxy voting record)

N-PX (Rule 30b1-4) is the annual fund-proxy-voting record for the twelve months ending June 30 (extended in 2022 to certain institutional managers' Section 14A say-on-pay votes). It is vote-by-vote structured data with no financial statements, no schedule of investments, and no SOX certifications. Overlap with N-CSRS is limited to filer population. N-PX answers a stewardship question; N-CSRS answers a financial-reporting and portfolio-composition question.

Boundary summary

N-CSRS Files is the only source that combines, in a single certified accession: (1) unaudited mid-year financial statements of registered management investment companies in as-filed shareholder-report form, with notes and accounting policies; (2) SOX 302/906 officer certifications attached to a semi-annual period; and (3) the schedule of investments as presented to shareholders alongside those financials. N-PORT delivers richer holdings more frequently but strips out everything else. N-CSR covers the same content shell for the annual cycle only. N-Q and N-30D are historical-continuity artifacts. N-MFP and N-PX address adjacent regimes. For the semi-annual certified shareholder report as a unified narrative-plus-financial document, N-CSRS Files is the sole direct source.

Who Uses This Dataset

The roles below depend on mid-year, fund-level disclosure that N-PORT and N-Q substitutes do not fully replace.

Competitive intelligence and product teams at fund complexes

Product strategists pull peer schedules of investments to compare holdings overlap, sector tilts, and position sizing; the statement of operations and notes for effective expense ratios, fee waivers, recoupment language, and 12b-1 economics; and the portfolio manager letter for attribution and positioning narrative. Output: competitive packs feeding fee, share-class, and new-fund decisions.

Compliance and internal audit at fund complexes

Compliance and disclosure-controls teams benchmark peer SOX 302 and 906 certifications, Item 4 disclosure-controls conclusions, and Rule 30a-3 procedure language. They mine phrasing on scope, exceptions, ICFR changes, and disclosure committee process to update certification templates and sub-certification programs.

Sell-side and buy-side equity analysts

Analysts use the schedule of investments to triangulate fund holdings between N-PORT publication windows, especially for private placements, restricted securities, and Rule 144A debt. They map fund-level concentration in covered names, flag rotations in or out, and use the statement of changes in net assets for share-class flow estimates. Output: ownership-overhang notes and squeeze-risk write-ups.

Plaintiffs' securities and class-action attorneys

Counsel pursuing Section 36(b) excessive-fee, style-drift, and disclosure claims compare advisory and sub-advisory fees, waivers, and expenses across the class period from the statement of operations and notes; test stated strategy against actual exposures (derivatives, leverage, illiquid positions) using the schedule of investments; and anchor officer-level representations to the SOX 302 and 906 certifications. Output: complaint exhibits, expert-report inputs, and damages-period reconstruction.

Academic researchers in finance, accounting, and governance

Researchers build longitudinal panels keyed on CIK, series and class identifiers, period of report, and accession number, joined to extracted line items and holdings. Common projects: fee competition, style drift, fair-value Level 1/2/3 classification, derivatives use, and disclosure tone in semi-annual letters.

Independent fund directors and their counsel

Trustees and independent counsel use peer N-CSRS filings during Section 15(c) advisory contract review to compare advisory fee schedules, breakpoints, waivers, and expense ratios across complexes with similar mandates. Notes and MD&A commentary inform questions on leverage, derivatives, and performance. Output: 15(c) board books and Gartenberg reasonableness memoranda.

Fund administrators and outsourced service providers

Administrators benchmark how peers categorize the schedule of investments, format the statement of assets and liabilities, and present derivatives, securities-lending collateral, and fair-value hierarchy tables. Output: onboarding playbooks, internal style guides, and pitch decks for administration mandates.

Risk and concentration teams at wealth platforms

Risk teams aggregate the schedule of investments across funds held by clients to compute look-through exposure to issuers, sectors, countries, and credit tiers, refreshing positions between N-PORT windows. Output: concentration dashboards, single-name advisor alerts, and rebalancing recommendations on look-through exposures.

Manager research at allocators and consultants

Insurance general accounts, pension plans, and investment consultants reconcile reported AUM against the statement of assets and liabilities, verify expense ratios and waivers against the statement of operations, and audit the schedule of investments for mandate consistency, leverage, and securities-lending exposure. Output: investment committee memoranda and watch-list decisions.

Data vendors building fund-financials feeds

Engineering teams parse the statement of assets and liabilities, statement of operations, statement of changes in net assets, financial highlights, and schedule of investments into normalized schemas keyed on CIK, series, and class, with period of report, filing date, and N-CSRS/A amendment flags anchoring versioning. Output: commercial fee databases, holdings panels, and screening feeds.

LLM and RAG developers for asset and wealth management

Teams use HTML and PDF primary documents, exhibits, and TXT renderings as training and evaluation material for fund-disclosure question answering. Workflows include retrieval over portfolio manager letters for advisor assistants, automated extraction of fees and expense ratios, and hallucination benchmarks on fund-specific facts.

Synthesis

The dataset serves fund-complex insiders, market-side analysts, plaintiffs' counsel, academics, fund-services providers, allocators, data vendors, and AI engineering teams. What unites them is the need for mid-year, certified, fund-level financial and holdings disclosure with consistent structure across two decades of EDGAR filings, accessed at scale.

Specific Use Cases

The following workflows are run directly against accession folders, the inline-XBRL primary report, the SOX certifications, and metadata.json.

1. Mid-year holdings refresh between N-PORT public-disclosure windows

Buy-side analysts and wealth-platform risk teams parse the schedule of investments from Item 7 of the primary HTML to refresh fund-level positions during the months when no public N-PORT report is available. They join seriesAndClassesContractsInformation[*] and entities[*].cik from metadata.json to client holdings, then aggregate CUSIP-level fair values across funds to produce look-through issuer, sector, and country exposures. Output: concentration dashboards and single-name advisor alerts keyed to periodOfReport.

2. Section 15(c) advisory-contract review packs

Independent fund directors and their counsel pull peer N-CSRS filings of comparable mandates and extract advisory and sub-advisory fees, breakpoints, waivers, and recoupment language from the statement of operations and notes (Item 7), plus expense ratios from the financial highlights. They also lift the board's basis-for-approval statement (Items 8–10 for open-end, Item 11 for closed-end) when present. Output: 15(c) board books and Gartenberg reasonableness memoranda.

3. Section 36(b) excessive-fee complaint construction

Plaintiffs' counsel reconstruct fee histories across the class period by extracting advisory fees, waivers, and expenses from Item 7 financials over consecutive N-CSRS and N-CSR accessions for the same CIK and series. They cross-check stated investment strategy in the Item 1 shareholder report against actual exposures (derivatives, leverage, illiquid Level 3 holdings) in the schedule of investments, and anchor officer-level representations to the EX-99.CERT and EX-99.906CERT signature lines. Output: complaint exhibits and damages-period reconstructions.

4. Inline-XBRL fund-financials feed for data vendors

Engineering teams at fund-data vendors run iXBRL-aware parsers across the primary HTML to extract oef:-tagged facts dimensioned along oef:ClassAxis, normalizing per-class expense ratios, total returns, and concentration percentages keyed on CIK, Series ID, Contract ID, and periodOfReport. They use the formType "N-CSRS/A" flag plus accessionNo/filedAt ordering to version-control restated figures. Output: commercial expense and holdings panels with class-level granularity.

5. SOX 302/906 certification benchmarking for fund-complex compliance teams

Compliance and disclosure-controls teams at fund complexes pull the EX-99.CERT and EX-99.906CERT exhibits across peer registrants of similar size and complexity, normalize the SGML wrapper, and diff the five-paragraph Section 302 attestation language for non-standard scope qualifiers, ICFR-change disclosures, and disclosure-committee references. They cross-reference Item 16 (Controls and Procedures) language in the primary report. Output: updated certification templates, sub-certification programs, and quarterly disclosure-committee training material.

6. Tailored Shareholder Reports compliance monitoring

Fund administrators and outside counsel track which open-end registrants have transitioned to the 2022 TSR format by detecting the structural split between a concise per-class TSR primary document under Item 1 and the full financial statements relocated to Item 7 (or to a separate EX-99 exhibit). They use documentFormatFiles[*].type and per-class oef:ClassAxis contexts to validate that each share class has a corresponding TSR. Output: TSR-readiness dashboards and exception lists for late or non-compliant series.

7. Closed-end fund buyback and corporate-action tracking

Closed-end fund analysts mine Item 14 (purchases of equity securities by the closed-end fund and affiliated purchasers) across the primary HTML for monthly share-repurchase tables, joining to discount/premium data from third-party sources. They also pull Item 15 disclosures on shareholder votes when present. Output: discount-management activism screens and buyback-cadence write-ups for closed-end coverage notes.

8. RAG corpora for fund-disclosure question answering

LLM teams building advisor-assistant and portfolio-manager copilots ingest the primary HTML (with SGML wrappers stripped where present), EX-99 attachments, and the metadata.json spine as a retrieval corpus, chunking by Item heading. They use entities[*].cik, seriesAndClassesContractsInformation, and periodOfReport as filter facets, and build evaluation sets of fund-specific factual questions (expense ratios, top-ten holdings, derivative usage) with answers grounded in tagged iXBRL facts. Output: production retrieval indexes and hallucination benchmarks scoped to a specific fund and reporting period.

Dataset Access

The Form N-CSRS Files Dataset is available through three access methods: a JSON metadata endpoint, a full dataset archive download, and per-container ZIP downloads. Filings cover the period from July 2003 to the present and are organized into monthly ZIP containers that include all documents from each EDGAR submission (excluding image files), along with a metadata file per accession number.

Dataset Index JSON API: https://api.sec-api.io/datasets/form-ncsrs-files.json

This endpoint returns dataset-level metadata (name, description, last updated timestamp, earliest sample date, total records, total size, covered form types, container format, and file types) along with the full list of available container files. Each container entry includes its key, size, record count, last updated timestamp, and a direct download URL. This endpoint does not require an API key. It can be polled regularly to detect which containers were refreshed in the latest run, so downstream pipelines can selectively re-download only the containers that changed.

          1
          {
        

          2
            "datasetId": "1f13365b-9ae0-68f9-a3fa-059a4eb256ad",
        

          3
            "datasetDownloadUrl": "https://api.sec-api.io/datasets/form-ncsrs-files.zip",
        

          4
            "name": "Form N-CSRS Files Dataset",
        

          5
            "updatedAt": "2026-04-28T03:03:29.872Z",
        

          6
            "earliestSampleDate": "2003-07-01",
        

          7
            "totalRecords": 222343,
        

          8
            "totalSize": 12894374908,
        

          9
            "formTypes": ["N-CSRS", "N-CSRS/A"],
        

          10
            "containerFormat": "ZIP",
        

          11
            "fileTypes": ["TXT", "JSON", "HTML", "PDF"],
        

          12
            "containers": [
        

          13
              {
        

          14
                "downloadUrl": "https://api.sec-api.io/datasets/form-ncsrs-files/2025/2025-10.zip",
        

          15
                "key": "2025/2025-10.zip",
        

          16
                "size": 142839201,
        

          17
                "records": 1583,
        

          18
                "updatedAt": "2025-11-02T04:18:42.000Z"
        

          19
              }
        

          20
            ]
        

          21
          }
        

Download Entire Dataset: https://api.sec-api.io/datasets/form-ncsrs-files.zip?token=YOUR_API_KEY

Downloads the complete archive containing all N-CSRS and N-CSRS/A filings from July 2003 onward in a single ZIP file. This endpoint requires an API key.

Download Single Container: https://api.sec-api.io/datasets/form-ncsrs-files/2025/2025-10.zip?token=YOUR_API_KEY

Downloads one monthly container ZIP file, which contains all filings submitted in that month. Use the container keys returned by the dataset index endpoint to construct the URL for any specific month. This endpoint requires an API key.

Frequently Asked Questions

What forms does this dataset cover?

The Form N-CSRS Files dataset covers Form N-CSRS (the certified semi-annual shareholder report) and Form N-CSRS/A (amendments to a previously filed N-CSRS). Both form types are filed by registered management investment companies under Section 30(b)(2) of the Investment Company Act of 1940 and Rule 30e-1 thereunder.

What does one record in this dataset represent?

One record is a single EDGAR accession of Form N-CSRS or N-CSRS/A, materialized on disk as an accession-numbered folder containing exactly one metadata.json, the primary report document, the Sarbanes-Oxley Section 302 (EX-99.CERT) and Section 906 (EX-99.906 CERT) certification exhibits, and any optional EX-99 attachments the registrant submitted. A single record may cover one fund or several series of a multi-series investment company filing one report at once.

Who is required to file Form N-CSRS?

Registered management investment companies — open-end mutual funds (including ETFs organized as open-end funds), closed-end funds, interval funds, and tender-offer funds — file N-CSRS. The legal filer is the registrant itself, and the principal executive officer and principal financial officer sign the SOX certifications. Unit investment trusts, business development companies, face-amount certificate companies, foreign funds not registered under the 1940 Act, and private funds relying on Section 3(c)(1) or 3(c)(7) do not file N-CSRS.

When is Form N-CSRS filed?

Form N-CSRS must be filed within 10 days after the registrant transmits the Rule 30e-1 semi-annual report to its stockholders. The clock runs from transmission, not from the close of the semi-annual period. Because each registrant (and each series) may have a different fiscal year, N-CSRS filings appear continuously throughout the calendar year.

What time period does the dataset cover?

The dataset begins with filings from 2003-07-01 — corresponding to the SEC's adoption of N-CSRS as part of its Sarbanes-Oxley implementation for registered investment companies — and is refreshed continuously as new accessions arrive on EDGAR.

What file format is the dataset distributed in?

The dataset is delivered as monthly ZIP containers organized as <year>/<year>-<month>.zip. Inside each container, records sit as flat sibling folders named with the bare 18-digit accession number. File types within a record include TXT, JSON, HTML, and PDF: modern filings are dominated by HTML primary reports and HTML certification exhibits plus the JSON metadata spine, while legacy filings may include .txt primaries or PDF supplements.

How does this dataset differ from Form N-PORT?

N-PORT (Rule 30b1-9) is a monthly, structured-XML portfolio holdings report with the third-month report of each fiscal quarter publicly disclosed after a 60-day lag, carrying granular position-level fields like ISIN/CUSIP/LEI, fair value, percent of net assets, and derivatives reference data. N-CSRS does not include those structured holdings fields but provides what N-PORT omits: the management discussion of performance, the statement of assets and liabilities, statement of operations, statement of changes in net assets, financial highlights, notes to financial statements, and SOX 302/906 officer certifications. The two are complements — N-PORT for granular holdings, N-CSRS for fund-level certified financial reporting.