The Form X-17A-5 Files Dataset is the corpus of annual audited reports filed on EDGAR by SEC-registered broker-dealers under Rule 17a-5(d) of the Securities Exchange Act of 1934, plus the corresponding amendments filed as Form X-17A-5/A. One record corresponds to a single EDGAR submission, identified by accession number, and bundles the EDGAR submission-envelope metadata, the structured XML facing page filed under the EDGAR seventeenafiler schema, the SEC's XHTML rendering of that facing page, and the filer-prepared PDF audit report(s) attached to the submission. The filer of record is the broker-dealer itself; the audit is performed by a PCAOB-registered independent public accountant named on the facing page. EDGAR digital coverage of these filings begins in November 2001 and continues to the present, with each accession packaged as a ZIP container alongside other monthly accessions.
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:
This dataset isolates one regulatory artifact: the annual audited Part III filing made by registered broker-dealers under Rule 17a-5(d), together with X-17A-5/A amendments. Form X-17A-5 is the annual audited report a broker-dealer registered under Section 15 of the Exchange Act must file no later than sixty calendar days after the end of its fiscal year. Each EDGAR accession is materialized as a folder named with the 18-digit accession number with dashes stripped (for example, 0001993563-25-000004 becomes 000199356325000004).
Each record folder bundles the complete public payload the broker-dealer submitted on EDGAR — primary report document, financial-statement attachments, and any supporting exhibits — minus image files and the SGML .txt submission wrapper, which is referenced by URL in the metadata rather than inlined. The dataset is distributed as monthly ZIP containers covering form types X-17A-5 and X-17A-5/A, with each accession folder carrying files in XML, PDF, JSON, TXT-link, and HTML form. Coverage begins on 2001-11-01, the start of EDGAR digital coverage for this form; earlier annual audited reports exist only in paper form in SEC and SRO archives.
The Form X-17A-5 submission has two distinct layers, both of which appear inside every record folder:
Form X-17A-5/A filings share the internal anatomy of X-17A-5 filings but represent amendments — typically a corrected facing page, a re-audited financial report, or a revised oath — to a previously accepted annual audited report.
Every record folder contains, at minimum, the following four artifacts, in this logical order:
metadata.json — a single-object JSON file describing the EDGAR submission envelope (form type, accession, filer entity, document inventory, links back to EDGAR-hosted copies).primary_doc.xml — the structured filing data, an <edgarSubmission> document under the http://www.sec.gov/edgar/seventeenafiler namespace, carrying every facing-page field plus the oath/affirmation.xslX-17A-5_X01/primary_doc.xml — a presentation copy of the same structured data, rendered as XHTML 1.0 Strict via the EDGAR XSL stylesheet (despite its .xml extension).Flow2025short.pdf, Harvestons.pdf, auditreportniadvisors630251.pdf). The single-PDF case dominates; a minority of filings carry two PDFs that split a short cover/facing-page report from the full financial-statements package (e.g. immg.pdf paired with immg4.pdf).The original .txt SGML wrapper that EDGAR rolls every submission into is not inlined into the folder; metadata.json instead carries linkToTxt pointing to the SEC-hosted copy. Image attachments referenced inside the original submission are also excluded from the folder; everything else the filer attached is preserved.
metadata.json is a flat JSON object (no top-level wrapper key) that gives a machine-readable index of the record: what form was filed, by whom, when, covering what fiscal period, and which files belong to it.
Key fields:
formType — X-17A-5 for an original annual audited report, X-17A-5/A for an amendment.accessionNo — the 18-digit dashed EDGAR accession; matches the folder name once dashes are added back.id — a 32-character hex internal identifier distinct from the accession.filedAt — ISO 8601 timestamp with timezone offset (-04:00 or -05:00) recording EDGAR receipt time.effectivenessDate — ISO date of effectiveness; optional, and absent on a meaningful portion of records.periodOfReport — fiscal year-end the audit covers, in ISO YYYY-MM-DD. Common values include December 31, June 30, and August 31, reflecting the variety of broker-dealer fiscal years.description — typically the literal string Form X-17A-5 - FOCUS Report.linkToTxt, linkToHtml, linkToFilingDetails — absolute URLs back to the EDGAR Archives copies. linkToFilingDetails resolves to the XSL-rendered XHTML facing page.linkToXbrl — present in the schema but always empty for this form type.documentFormatFiles[] — array enumerating every file the filer attached on EDGAR. Each element carries sequence (string), size (byte count as string), documentUrl, type (commonly X-17A-5 for the structured submission, FULL for the audit PDF, sometimes a single space " " for the rolled-up .txt), and an optional description such as "PUBLIC", "AUDIT REPORT 08312025", "FYE 06-30-2025 FINANCIAL STATEMENTS & NOTES", or "Complete submission text file".entities[] — a one-element array describing the broker-dealer filer. Fields include companyName (suffixed with (Filer)), cik, irsNo, fileNo (the SEC 008- series broker-dealer file number), fiscalYearEnd (MMDD string, e.g. "0831"), stateOfIncorporation (two-letter code), act (always "34", the Exchange Act), type ("X-17A-5"), and filmNo (8-digit EDGAR film number; optional, absent when the film number has not yet been assigned at extraction time).seriesAndClassesContractsInformation — always [] for this form type.dataFiles — always [] (no XBRL data files are attached).Format quirks worth noting for downstream extraction: numeric-looking fields (size, sequence, cik, irsNo, filmNo, fiscalYearEnd) are encoded as strings rather than numbers, and several fields use a single literal space character " " to indicate "not applicable / unknown" rather than null or empty string.
The structured submission is rooted at <edgarSubmission xmlns="http://www.sec.gov/edgar/seventeenafiler" xmlns:com="http://www.sec.gov/edgar/common"> and contains exactly two children, <headerData> and <formData>. Together these capture the full content of the regulatory facing page.
<headerData> carries the EDGAR-side filer credentials and submission flags:
<submissionType> — X-17A-5 or X-17A-5/A.<filerInfo>/<filer>/<filerCredentials> — <filerCik> (10-digit zero-padded CIK) and <filerCcc>. The CCC is always masked to the literal XXXXXXXX in the public archive; it is a redaction, not real credential material.<filerInfo>/<filer>/<confirming008CopyNumber> — optional, present only when the filing is a confirming copy directed at a specific broker-dealer file number (e.g. 008-46293).<liveTestFlag> — LIVE or TEST.<flags> — three booleans: returnCopyFlag, confirmingCopyFlag, overrideInternetFlag.<formData> carries the substantive facing-page disclosures:
<submissionInformation> — <periodBegin> and <periodEnd> formatted in US-ordered MM-DD-YYYY (note the format mismatch with metadata.json, which uses ISO YYYY-MM-DD); <typeOfRegistrant>/<typeOfBDRegistrant> taking one of Broker-dealer, Security-based swap dealer, Major security-based swap participant, or OTC derivatives dealer; and <materialWeakness> (Y or N), the registrant's declaration of whether the audit identified a material weakness in internal control over financial reporting.<registrantIdentification> — the broker-dealer's facing-page identification block: <brokerDealerName>, <businessAddress> (with com:-namespaced children street1, optional street2, city, stateOrCountry, zipCode), <contactPersonName>, and <contactPersonPhoneNumber>.<accountantIdentification> — <accountantName> (the audit firm), <accountantAddress> (same com: address schema), and <accountantType> taking one of Certified Public Accountant or Certified Public Accountant not resident in United States or any of its possessions.<oathSignature> — the sworn oath or affirmation: <signPersonName>, <entityName>, <signDate> (MM-DD-YYYY), an optional <explanation> element holding free-text exception language (most often the literal NO EXCEPTIONS), <signature>, <oathTitle> capturing the signer's role at the broker-dealer (e.g. CEO and President, CFO), and <confirmNotarizedFlag> (Y/N) indicating whether the oath was notarized.This XML body is the canonical machine-readable source for every facing-page field and is the file most useful for structured extraction.
The subfolder xslX-17A-5_X01/ contains a single file, primary_doc.xml, which despite the .xml extension is an XHTML 1.0 Strict document (declared with <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" ...>) linking /css/SDR_print.css. It is the EDGAR-rendered presentation copy of the facing page, mirroring every field of the sibling structured XML — CIK, masked CCC, period, registrant address, accountant address, oath text, notarization flag — through the standard EDGAR XSL stylesheet. It contains no information not already present in the structured primary_doc.xml; its role is purely human display, and metadata.json's linkToFilingDetails resolves to its EDGAR-hosted equivalent.
The substantive content of the filing — the work product the broker-dealer is actually filing for — lives in one (occasionally two) PDFs at the root of the accession folder. These are filer-prepared documents with no enforced naming, slug, or layout convention; the filename is whatever the registrant chose at submission. Sizes typically range from roughly 2 MB to over 10 MB.
Each PDF is the public version of the audited financial report and ordinarily contains, in narrative-and-tabular form:
When two PDFs are attached, they typically split into a short cover-page report and a longer financial-statements package; the only reliable signal distinguishing the two within the record is the description field on the corresponding entry in documentFormatFiles[] inside metadata.json. Filer-chosen filenames carry no schema.
Broker-dealers may also file a confidential supplement to the audit report directly with the SEC outside of EDGAR; that supplement is never received by EDGAR and is therefore not present in the dataset.
Each record folder contains:
metadata.json);primary_doc.xml);xslX-17A-5_X01/primary_doc.xml);This covers the complete public payload the broker-dealer submitted on EDGAR — primary report document, financial-statement attachments, and any supporting exhibits — minus the items called out below.
The following are not contained inside the record folder:
.txt SGML submission wrapper that EDGAR generates from the filer's package. It is not inlined; metadata.json carries linkToTxt pointing at the SEC-hosted copy instead.<filerCcc> element is always redacted to XXXXXXXX in the public archive.metadata.json uses ISO 8601 (YYYY-MM-DD) for periodOfReport and effectivenessDate, while primary_doc.xml uses US-ordered MM-DD-YYYY for <periodBegin>, <periodEnd>, and <signDate>. Any join across the two layers must normalize date formats.0001993563-25-000004 becomes 000199356325000004); metadata.json.accessionNo retains the dashes.effectivenessDate, entities[0].filmNo, <confirming008CopyNumber>, and <oathSignature>/<explanation> are not always present. Their absence is not a data error.metadata.json (size, sequence, cik, irsNo, filmNo, fiscalYearEnd) are encoded as strings, and a single literal space " " is used in some string fields (sequence, size, type) to mean "not applicable", rather than null or empty string.<filerCcc> value is always the literal XXXXXXXX and must not be treated as a credential.<confirmingCopyFlag> set true and a populated <confirming008CopyNumber>. These are duplicate filings directed at a specific 008-series broker-dealer file number rather than independent annual audit submissions.<submissionType> (and metadata.json.formType) is X-17A-5/A, and the attached PDFs typically supersede or correct the corresponding documents in an earlier accession by the same registrant for the same periodOfReport. The dataset does not link amendment chains explicitly; pairing an original with its amendment requires joining on entities[0].cik and periodOfReport.description text on the corresponding documentFormatFiles[] entries is the only reliable way to distinguish them.linkToXbrl is empty and dataFiles is [] on every record.<materialWeakness> in <submissionInformation> is the registrant's facing-page declaration of whether the audit identified a material weakness in internal control over financial reporting; it is a single Y/N field and does not capture the substantive description, which lives only in the PDF.<typeOfBDRegistrant> accommodates security-based swap dealers, major security-based swap participants, and OTC derivatives dealers in addition to ordinary broker-dealers, reflecting Rule 17a-5's expanded scope under Title VII of the Dodd-Frank Act.Form X-17A-5 is filed by the broker-dealer registrant itself. The reporting universe is defined by SEC registration, not by public-company status:
The independent public accountant who audits the financial statements is named on the facing page and signs the audit, compliance, or exemption reports, but is not the filer of record. The auditor must be registered with the Public Company Accounting Oversight Board (PCAOB) — a requirement added by the SEC's 2013 amendments to Rule 17a-5, which also introduced the compliance-report and exemption-report regime and aligned broker-dealer audits with PCAOB auditing standards. The principal officer, general partner, manager, or member who signs the oath or affirmation signs on behalf of the registrant, not in a personal filer capacity.
The filing is periodic and fiscal-year-driven. The trigger is the close of the broker-dealer's fiscal year.
Because the obligation is keyed to broker-dealer registration under Section 15 or 15C, several adjacent populations are outside the dataset:
A record exists in this dataset only if the entity was a registered broker-dealer during the relevant fiscal year. Customers, counterparties, and affiliates referenced in supplemental schedules or notes are subjects of disclosure, not filers.
Several adjacent forms cover the same filer population or share an "annual audited" feel, but each occupies a distinct slot. The comparisons below sharpen those boundaries.
Same form family, same rule (17a-5), same filers, but a different report. Parts I (monthly, carrying firms), II (quarterly, clearing/carrying firms), and IIA (quarterly, introducing firms) are unaudited operational and financial snapshots used for ongoing SROs and SEC surveillance. This dataset covers only Part III, the once-a-year auditor-attested package due 60 days after fiscal year-end. Parts I/II/IIA are largely confidential regulatory submissions; Part III public portions are the EDGAR-visible annual audit.
Same form, same filers; corrects or supersedes a prior Part III submission and is included in this dataset alongside originals.
Status and identity filings, not financial reports. Form BD registers a broker-dealer; BD-N is a notice filing for certain bank-affiliated entities; BD-W withdraws registration. Content is ownership, control persons, disciplinary history, and business activity disclosures. Most BD-family content lives in FINRA's CRD/IARD systems and is exposed via BrokerCheck rather than EDGAR.
Filed by registered investment advisers under Advisers Act Rule 206(4)-2 to describe custodial arrangements over client assets. Overlaps thematically because broker-dealers often act as qualified custodians, and the X-17A-5 audit includes a Rule 15c3-3 customer-protection compliance or exemption report. Different statute (Advisers Act vs. Exchange Act), different filer population (RIAs vs. BDs), different content (descriptive custodial relationships vs. audited financials and 15c3-3 compliance).
SIPC-administered membership and assessment filings tied to the same broker-dealer population. They record SIPC membership status and net operating revenue-based assessments, not audited financial position. Narrow, formulaic, and largely outside the EDGAR X-17A-5 corpus.
Filed by larger broker-dealers under Rules 17h-1T/17h-2T to disclose risk concentrations at Material Associated Persons (parent and affiliate entities). Closer in spirit to X-17A-5 than the registration forms, but the entity perimeter and access regime differ: Form 17-H covers the affiliates and parent, X-17A-5 covers the broker-dealer itself. 17-H is largely confidential; X-17A-5 public portions are openly filed.
The most easily confused comparison: both are "annual" and both contain audited financial statements. They are otherwise very different instruments.
Most broker-dealers are private subsidiaries and never file a 10-K; most 10-K filers are not broker-dealers. When a broker-dealer is itself public, both can exist for the same year and contain genuinely different content.
This dataset is a curated subset of the EDGAR firehose, filtered to form types X-17A-5 and X-17A-5/A, with each accession packaged as a ZIP container containing a metadata file plus all original submission documents (XML, PDF, JSON, TXT, HTML), excluding images. Versus the raw archive, it skips the work of filtering down to the broker-dealer audit packages while preserving original accession structure.
This dataset isolates one filing type: the annual audited Part III report under Rule 17a-5(d), with amendments, for the registered broker-dealer population on EDGAR. It is not a surveillance feed (FOCUS I/II/IIA), not a registration or status record (BD family), not a custody description (Form Custody), not a SIPC assessment record, not a group-level risk file (17-H), not an issuer-level public report (10-K), and not the unfiltered EDGAR archive. Its value is delivering the one consistently audited, broker-dealer-entity-level disclosure of financial condition and customer-protection compliance as a single packaged corpus.
Different professionals reach for different layers of the X-17A-5 package: the primary_doc.xml metadata, the statement of financial condition, the supplemental net-capital and Rule 15c3-3 schedules, the independent accountant's report, and the signed oath or affirmation.
Series 27/28 principals, CFOs, and compliance staff at registered broker-dealers benchmark their own filings against peers. They read the statement of financial condition, the net capital computation, and the customer reserve / Rule 15c3-3 schedules to compare presentation of subordinated liabilities, allowable assets, haircuts, and PAB inputs. The corpus supports pre-filing review, disclosure training, and reconciliation against internal FOCUS reports.
Examination and risk-monitoring staff use the historical corpus to track registrant financial condition across fiscal years, flag outliers in net capital, and surface unusual movement in subordinated debt, related-party receivables, or off-balance-sheet items. They focus on the audited balance sheet, capital-structure notes, supplemental net-capital schedules, and any modifications, emphasis-of-matter, or going-concern paragraphs in the accountant's report. Outputs feed exam scoping and supervisory follow-up where audited figures diverge from monthly FOCUS data.
Engagement partners and quality-control reviewers use prior-year filings for client acceptance, continuance, and peer review. They examine the predecessor auditor's opinion, the supplemental report on internal control over compliance, and disclosures distinguishing exemption provisions from full Rule 15c3-3 compliance. Inspection-readiness teams study opinion wording and supplemental-report structure across comparable engagements.
Credit officers and counterparty-risk analysts at custodian banks, prime brokers, clearing firms, and tri-party agents size exposures to non-public broker-dealer counterparties. They pull the statement of financial condition and notes on collateral, securities borrowed and loaned, repurchase agreements, fair-value hierarchy, and credit-risk concentrations. The audited balance sheet anchors internal ratings, line sizing, margin terms, KYC files, and annual credit reviews.
Securities litigators, forensic accountants, and trustees handling arbitrations, insolvencies, and SIPA liquidations reconstruct the trajectory of distressed broker-dealers from the historical sequence. They focus on year-over-year net capital, subordinated loan agreements, contingent liabilities, related-party notes, going-concern language, and amendments filed as X-17A-5/A. The signed oath or affirmation page is load-bearing in matters alleging certification misstatements.
Researchers in financial economics and capital-adequacy policy build longitudinal panels of broker-dealer condition. They extract structured fields from primary_doc.xml (registrant identifier, fiscal-year-end, auditor, filing date), then parse the audited balance sheet and income statement for measures of leverage, asset composition, and capital cushion. Studies cover post-crisis balance-sheet evolution, audit-market concentration, and the effect of rule changes on net capital.
Bankers and corporate development staff evaluating broker-dealer acquisitions use X-17A-5 as the foundational diligence document for private targets. They review the audited statement of financial condition, capital-structure notes, the auditor's report, and the multi-year history to assess earnings quality, seasonality, and any net capital deficiencies. The dataset supports valuation, treatment of subordinated debt that counts toward net capital, and change-in-control planning.
Quantitative modelers building PD/LGD and stress frameworks for broker-dealer counterparties standardize balance-sheet line items across registrants and years, anchoring identifiers on the metadata XML and treating X-17A-5/A amendments as restatement events. The audited PDF and JSON-normalized records provide the time-series consistency that FOCUS-only signals lack.
Reporters at financial trade publications use the dataset to access audited financials of firms with no other public reporting. They look at the facing page (registrant and accountant identity), the opinion, legal-proceedings and related-party notes, and year-over-year changes ahead of enforcement, customer-harm, or insolvency stories.
Engineering teams building retrieval systems for compliance, diligence, and audit-quality workflows ingest the corpus across PDF (document fidelity), HTML and TXT (search), XML (metadata), and JSON (normalized records). They train balance-sheet extractors, ground question-answering over auditor reports, and back domain assistants with real broker-dealer disclosure.
The X-17A-5 corpus supports a range of concrete workflows, each anchored to specific parts of the record.
Credit officers at custodian banks, prime brokers, and tri-party agents pull the statement of financial condition and the related notes on securities borrowed/loaned, repurchase agreements, collateral, and fair-value hierarchy from the audit PDF, then key the record by entities[0].cik and periodOfReport from metadata.json. The output feeds internal ratings, line sizing, margin add-ons, and the annual credit review file for private broker-dealer counterparties that have no 10-K.
Series 27/28 principals and broker-dealer CFOs extract the supplemental Rule 17a-5 schedules from peer audit PDFs (computation of net capital, computation for determination of reserve requirements, information relating to possession or control) and reconcile presentation of allowable assets, haircuts, subordinated liabilities counted toward net capital, and PAB inputs against their own draft FOCUS-derived figures. Pre-filing reviews use the corpus to flag presentation gaps before submission.
Risk-monitoring staff diff the audited balance sheet and net-capital computation against monthly FOCUS Part II/IIA data for the same registrant, then read the accountant's report and the supplemental compliance or exemption report for emphasis-of-matter, going-concern, or modified-opinion language. Registrants that show divergence between audited and surveillance figures, a <materialWeakness> flag of Y on the facing page, or a switched accountant in <accountantIdentification> get prioritized for exam follow-up.
Engagement partners pull prior-year filings for prospective clients, reading the predecessor auditor's opinion and the supplemental report on internal control over compliance to evaluate engagement risk and the exemption-versus-full-15c3-3 posture. Quality-control teams pivot on <accountantName> across the corpus to assemble peer engagements with comparable customer-protection scope for opinion-wording and supplemental-report structure benchmarking.
Litigators, forensic accountants, and SIPA trustees pair each X-17A-5/A amendment to its original by joining on entities[0].cik and periodOfReport, then diff the amended statement of financial condition, capital-structure notes, supplemental schedules, and the <oathSignature> block (signer name, title, notarization flag, free-text <explanation>) against the superseded filing. The signed oath or affirmation page is load-bearing in matters alleging certification misstatements or false net-capital reporting.
Researchers normalize structured fields from primary_doc.xml (CIK, fiscalYearEnd, <typeOfBDRegistrant>, <accountantName>, <materialWeakness>, <signDate>) into a panel keyed by registrant-year, then run PDF extraction over the statement of financial condition and supplemental net-capital schedule to build leverage, asset-composition, and capital-cushion series. Studies cover post-Dodd-Frank balance-sheet evolution, broker-dealer audit-market concentration (using <accountantName> and <accountantAddress>), and the impact of Rule 17a-5 amendments on customer-protection posture.
Corporate development teams evaluating broker-dealer acquisitions assemble three-to-five fiscal years of the target's audit PDFs as the foundational diligence pack, focusing on the statement of financial condition, capital-structure notes, the auditor's report, and any X-17A-5/A amendments that signal prior restatements. The supplemental net-capital schedule drives valuation adjustments for subordinated debt that does or does not count toward net capital under change-in-control.
The Form X-17A-5 Files Dataset is accessible through three endpoints: a public JSON index for metadata discovery, a full archive download, and per-container downloads for incremental retrieval.
Dataset Index JSON API: https://api.sec-api.io/datasets/form-x17a5-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), the download URL for the full archive, and a list of individual container files. Each container entry includes its key, size, record count, last updated timestamp, and direct download URL. Poll this endpoint to detect which monthly containers were refreshed in the latest run and download only the changed files. This endpoint does not require an API key.
Example response:
1
{
2
"datasetId": "1f13365b-9ae0-68ec-916a-3e0a7f9fc8f2",
3
"datasetDownloadUrl": "https://api.sec-api.io/datasets/form-x17a5-files.zip",
4
"name": "Form X-17A-5 Files Dataset",
5
"updatedAt": "2026-05-08T02:51:19.000Z",
6
"earliestSampleDate": "2001-11-01",
7
"totalRecords": 57169,
8
"totalSize": 37319466773,
9
"formTypes": ["X-17A-5", "X-17A-5/A"],
10
"containerFormat": "ZIP",
11
"fileTypes": ["XML", "PDF", "JSON", "TXT", "HTML"],
12
"containers": [
13
{
14
"downloadUrl": "https://api.sec-api.io/datasets/form-x17a5-files/2026/2026-04.zip",
15
"key": "2026/2026-04.zip",
16
"size": 13818783,
17
"records": 154,
18
"updatedAt": "2026-05-08T02:51:19.000Z"
19
}
20
]
21
}
Download Entire Dataset: https://api.sec-api.io/datasets/form-x17a5-files.zip?token=YOUR_API_KEY
Downloads the complete dataset as a single ZIP archive covering all filings from the earliest sample date (2001-11-01) onward. This endpoint requires an API key passed via the token query parameter.
Download Single Container: https://api.sec-api.io/datasets/form-x17a5-files/2026/2026-04.zip?token=YOUR_API_KEY
Downloads a single monthly container ZIP. Replace the <year>/<year>-<month>.zip path segment with any key returned by the dataset index JSON API to fetch a specific month. This endpoint requires an API key passed via the token query parameter.
The dataset covers Form X-17A-5 — the annual audited report a registered broker-dealer must file under Rule 17a-5(d) of the Securities Exchange Act of 1934 — together with Form X-17A-5/A, which is the amendment counterpart filed to correct or supersede a previously accepted annual report. It does not cover the monthly or quarterly FOCUS reports filed as Form X-17A-5 Part I, II, IIA, or IIB through SROs.
One record is a single EDGAR submission of Form X-17A-5 or X-17A-5/A, identified by EDGAR accession number. The record is materialized as a folder bundling metadata.json, the structured XML facing page (primary_doc.xml under the seventeenafiler namespace), the EDGAR XSL-rendered XHTML copy of that facing page, and the filer-prepared PDF audit report(s).
Every broker or dealer registered with the SEC under Section 15 of the Exchange Act must file Form X-17A-5, including introducing brokers, market makers, prime brokers, retail and institutional brokerage firms, and government securities brokers and dealers registered under Section 15C. Both carrying (clearing) broker-dealers and non-carrying broker-dealers file; the supplemental schedules and the choice between a compliance report and an exemption report differ between them.
The annual report must be filed no later than 60 calendar days after the end of the broker-dealer's fiscal year. The 60-day deadline is fixed and applies uniformly regardless of firm size, accelerated-filer status, or any Section 13 or 15(d) issuer reporting obligation. Amendments (X-17A-5/A) have no fixed deadline and are filed when an underlying error or revision is identified.
EDGAR digital coverage of Form X-17A-5 begins on 2001-11-01 and continues to the present, refreshed in monthly containers. Earlier annual audited reports filed before EDGAR digital coverage exist only in paper form in SEC and SRO archives and are not part of the dataset.
Each record folder includes the EDGAR submission-envelope metadata (metadata.json), the structured facing-page XML (primary_doc.xml), the XHTML rendering of that facing page (xslX-17A-5_X01/primary_doc.xml), and the public audit report PDF(s) attached by the filer, including any supporting exhibits the filer attached as additional PDFs. The original .txt SGML submission wrapper is referenced by URL in metadata.json but not inlined; image files are deliberately excluded; the confidential supplement to the audit report (filed outside EDGAR) is never received by EDGAR and is therefore not present; and the <filerCcc> value is always redacted to XXXXXXXX in the public archive.
Both are annual and both contain audited financial statements, but the rule, filer perimeter, scope, and structure differ. Form 10-K is filed by issuers with publicly traded securities under Sections 13/15(d), is comprehensive issuer disclosure, and is heavily XBRL-tagged. Form X-17A-5 is filed by registered broker-dealers under Rule 17a-5(d) regardless of whether the firm is a public issuer, focuses narrowly on the broker-dealer entity's net capital, customer protection, and related schedules, and consists predominantly of PDF audit reports with limited structured data.
The dataset is exposed through three endpoints at api.sec-api.io: a public JSON index at /datasets/form-x17a5-files.json (no API key required) for discovering containers and refresh timestamps; a full-archive ZIP download at /datasets/form-x17a5-files.zip?token=YOUR_API_KEY; and per-month container downloads at /datasets/form-x17a5-files/<year>/<year>-<month>.zip?token=YOUR_API_KEY for incremental retrieval. Containers are distributed as ZIP archives, and individual record files are XML, PDF, JSON, TXT-link, and HTML.