The Form S-B Files Dataset is a corpus of Schedule B registration statements filed on EDGAR by foreign sovereign governments, their political subdivisions, and supranational financial institutions registering securities offerings — almost always sovereign or sovereign-guaranteed debt — in the United States under the Securities Act of 1933. One record corresponds to a single accepted EDGAR filing (an original S-B or an S-B/A amendment), packaged as a per-accession folder containing a metadata.json index and the original-submission HTML, PDF, and TXT documents EDGAR received. Filers include sovereigns such as the State of Israel, the United Mexican States, and the Republic of Korea; political subdivisions such as the Province of Quebec; and supranationals such as the International Bank for Reconstruction and Development (IBRD), the Inter-American Development Bank, and the European Investment Bank. The dataset covers filings from January 2002 forward, the point at which the S-B form code became available in EDGAR's electronic filing system, and includes both initial registrations and pre- and post-effective amendments in the form types S-B and S-B/A.
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 Form S-B Files Dataset contains every Schedule B registration statement (Form S-B) and Schedule B amendment (Form S-B/A) accepted by EDGAR from January 2002 to the most recent refresh. Schedule B is the registration regime prescribed by Sections 7 and 8 and Schedule B of the Securities Act of 1933 for foreign governments, their political subdivisions, and supranational entities; it is a deliberately abbreviated disclosure regime — fiscal accounts, funded-debt schedules, securities terms, underwriter identification, and an opinion of counsel as to legality — used in lieu of the longer Schedule A regime that domestic corporate registrants follow on Forms S-1, S-3, S-4, and the like.
The dataset captures the full filer population for this regime: foreign sovereign issuers (the State of Israel, the United Mexican States, the Republic of Argentina, the Federative Republic of Brazil, the Republic of Chile, the Republic of the Philippines, the Republic of Indonesia, the Republic of Korea, the Republic of South Africa, and others), political subdivisions of foreign governments (principally Canadian provinces such as Quebec, Ontario, Manitoba, and British Columbia), supranational and multilateral financial institutions (IBRD/World Bank, the Inter-American Development Bank, the Asian Development Bank, the African Development Bank, the European Investment Bank, the European Bank for Reconstruction and Development), and agencies and instrumentalities of foreign governments where the obligation is sovereign in character. Issuer concentration is high and total record volume is modest relative to commercial-issuer datasets.
Records are bundled into ZIP containers organized by the calendar month of EDGAR acceptance, so the natural shard is YYYY-MM.zip and the natural row inside each shard is one accession-number subfolder. The file types in the dataset are HTML (the dominant document carrier), TXT (occasional ancillary artifacts), JSON (the per-record manifest), and PDF (used when an exhibit was delivered as PDF rather than HTML).
One record in the Form S-B Files Dataset corresponds to a single Schedule B filing accepted by EDGAR — either an original S-B registration statement or an S-B/A amendment. The record unit is one accession number: every non-image document EDGAR received under that accession is grouped into a single per-filing folder, accompanied by a dataset-side metadata.json index. Schedule B is the registration regime used by foreign sovereigns, their political subdivisions, and supranational entities, so a record almost always represents a sovereign-debt or sovereign-guaranteed offering by an issuer such as the State of Israel, the Province of Quebec, the United Mexican States, the Kingdom of Belgium, the European Investment Bank, or the African Development Bank.
Form S-B is the registration statement prescribed under Schedule B of the Securities Act of 1933 (Sections 7 and 8 and Schedule B itself). Schedule B is a deliberately abbreviated disclosure regime intended for sovereign and supranational issuers, in lieu of the longer Schedule A regime used by domestic corporate registrants on Forms S-1, S-3, S-4, and the like. Its statutory items include:
An S-B/A is a pre-effective amendment or post-effective amendment to a previously filed S-B, used to correct, supplement, or update the original registration statement. Amendments most often add pricing supplements, revise the Calculation of Registration Fee table, or attach further exhibits.
Inside each ZIP, every record is laid out as one folder whose name is the 18-digit accession number with dashes stripped (for example, accession 0001104659-25-071583 becomes folder 000110465925071583), nested under a top-level directory whose name mirrors the container's YYYY-MM, so an extracted path looks like 2025-07/000110465925071583/. The folder contains two structural layers:
metadata.json index that is always present and describes the filing, its parties, and every document EDGAR received; and.txt artifacts.Image files (.gif, .jpg, .png) delivered with the original submission are deliberately omitted from the ZIP even though they remain enumerated in metadata.json. The file types found in the dataset are HTML, TXT, JSON, and PDF: JSON is the per-record manifest, HTML is the dominant document carrier, and PDF appears only when a particular filing chose to deliver an exhibit or supplement as PDF rather than HTML.
metadata.json schemametadata.json is a single JSON object that names the filing, links it back to EDGAR, and enumerates its parties and documents. The keys carry the following meaning in the Schedule B context:
formType — the EDGAR form type for this accession, either S-B or S-B/A.accessionNo — the canonical 18-digit dashed accession number (e.g. 0001104659-25-071583); together with formType this uniquely identifies the record.linkToFilingDetails — absolute URL of the primary document on www.sec.gov/Archives/....description — EDGAR's human-readable filing description, typically Form S-B - Registration statement for securities of Foreign Governments and Subdivisions, with [Amend] appended for S-B/A.linkToTxt — URL of the EDGAR full-submission .txt envelope (the SGML-wrapped concatenation of every document in the submission).linkToHtml — URL of the EDGAR -index.htm filing index page.linkToXbrl — empty string for this dataset; Schedule B issuers are outside the structured-data tagging regime that applies to domestic corporate registrants.filedAt — ISO-8601 acceptance timestamp with timezone offset, e.g. 2025-07-29T12:18:13-04:00.id — a dataset-internal hexadecimal record identifier.documentFormatFiles — an array enumerating every document EDGAR received as part of the submission. Each element carries sequence (the EDGAR document sequence number, encoded as a string; the full-submission text file uses a single space), size in bytes (also string-encoded), documentUrl pointing to the file on EDGAR, description, and type (an EDGAR document-type label such as S-B, S-B/A, GRAPHIC, EX-99, or blank for the full-submission text envelope). This array is a manifest of what EDGAR received, not of what is present on disk inside the ZIP, because graphics and the redundant TXT envelope are not extracted.dataFiles — an array of machine-readable data files (XBRL/XML); empty for Schedule B records.seriesAndClassesContractsInformation — an array of investment-company series/class identifiers; empty for Schedule B records, since Schedule B is not an investment-company regime.entities — an array of one or more entity objects identifying the parties associated with the filing. For Schedule B records the relevant entity is almost always a single sovereign filer carrying companyName (e.g. ISRAEL, STATE OF (Filer)), cik (e.g. 52749 for the State of Israel), type matching the filing's formType, act set to 33 for the Securities Act of 1933, fileNo (the SEC file number, of the form 333-XXXXXX), filmNo (the EDGAR film number for the acceptance), sic (typically 8888 Foreign Governments), irsNo (usually 000000000, since foreign governments are not assigned U.S. employer identification numbers), and fiscalYearEnd encoded as MMDD.The fileNo field is structurally important for amendment tracking: an S-B/A reuses the original S-B's SEC file number, so the full amendment chain for a registration can be reconstructed by grouping records on fileNo together with cik.
The principal document inside each record folder is the registration statement itself, delivered with a .htm extension. Although the extension is HTML, the file as stored on disk opens with EDGAR's SGML submission wrapper rather than a bare <html> root: a <DOCUMENT> block prefixes the HTML body, declaring <TYPE> (S-B or S-B/A), <SEQUENCE> (1 for the primary document), <FILENAME>, <DESCRIPTION>, and a <TEXT> opener; after the closing </HTML> the wrapper continues with </TEXT></DOCUMENT>. Strict HTML parsers will trip over the leading SGML tags, so consumers should either strip the wrapper before parsing or use an SGML-tolerant reader.
Within the HTML body, the primary document is organized in the conventional Schedule B order:
For pricing-supplement-style amendments and post-effective amendments, the primary document may be substantially shorter, comprising only the new pricing terms, an updated Calculation of Registration Fee, and confirming legal opinions, while incorporating the bulk of the original registration statement by reference.
Each record contains every document EDGAR received under the accession that the dataset publisher chose to extract: the metadata.json index, the SGML-wrapped HTML primary document, and any non-image exhibits or supplements (additional HTML files, PDF attachments, occasional .txt artifacts). The documentFormatFiles array in metadata.json enumerates the complete EDGAR manifest, including items not extracted, so consumers always have a faithful inventory of the original submission even when the corresponding bytes are absent from disk.
Three classes of content are excluded from each record folder:
.gif, .jpg, .png) — graphics such as sovereign seals, signature scans, and branded prospectus covers are listed in documentFormatFiles (typically with type: GRAPHIC) but their bytes are not packaged into the ZIP. Filings can therefore appear "incomplete" relative to the EDGAR index by design.documentFormatFiles typically includes an entry whose sequence is a single space and whose description is Complete submission text file (e.g. 0001104659-25-071583.txt). This SGML envelope concatenates every document in the submission and is generally not extracted as a separate file, since its contents duplicate the individual documents that are extracted.dataFiles and seriesAndClassesContractsInformation are uniformly empty for this dataset. There is no XBRL, no XML financial-statement data, and no investment-company series-and-class information attached to Schedule B filings.Content packaged separately — the EDGAR filing-index HTML page itself, related correspondence, and any companion 18-K annual reports filed by the same sovereign — does not live inside the S-B record folder. It is reachable through linkToHtml and linkToFilingDetails in metadata.json but belongs to other datasets.
S-B/A versus S-B)S-B/A records are structurally indistinguishable from original S-B records at the packaging layer: identical folder layout, identical metadata.json schema, identical SGML-wrapped HTML primary document, identical exhibit conventions. The amendment status is encoded in three places: the top-level formType of metadata.json, the type field of each entity object inside entities, and the <TYPE> header inside the SGML wrapper of the primary document. EDGAR's description additionally appends the suffix [Amend] for amendments.
Amendment chains are reconstructed by joining records on the SEC fileNo (333-XXXXXX) carried in each entity, since pre-effective and post-effective amendments reuse the original registration's file number; the cik of the filer is invariant across the chain. Amendment numbering itself is not exposed as a separate metadata field — Amendment No. 1, Amendment No. 2, and so on are conventionally indicated only on the cover page of the primary document.
The earliest covered date is January 2002, several years after EDGAR's transition from ASCII/text-only to HTML-based filings. Every record in the dataset is therefore delivered in HTML (wrapped in the EDGAR SGML envelope) rather than flat ASCII; there is no plain-text era to accommodate inside the dataset window. Across the 2002-to-present span, structural conventions for Schedule B filings have been notably stable: the Calculation of Registration Fee table format, the Schedule B item ordering, the underwriter and counsel disclosure conventions, and the prospectus-supplement amendment pattern have changed only at the margin. The most visible drift is presentational — older filings use simpler HTML with fewer inline styles and heavier reliance on <TABLE> markup, while later filings use richer CSS, embedded font declarations, and more polished typesetting — but document order and disclosure content are consistent.
Container layout itself has not evolved within the dataset: the per-month ZIP, the YYYY-MM/<accession>/ folder shape, the always-present metadata.json, the dash-stripped accession-number folder name, and the omission of image files are uniform across years. PDF attachments occasionally appear as exhibits when an issuer chose that format for an opinion or fiscal agency agreement, but the primary registration document is consistently HTML throughout.
A handful of nuances matter for downstream extraction:
.htm documents will defeat strict HTML parsers; either strip everything before the first <HTML> and after the last </HTML> or use a tolerant parser.documentFormatFiles is a manifest of what EDGAR received, not of what is present on disk; consumers expecting every enumerated file to exist will find systematic "missing" image entries and a missing TXT envelope.size and sequence inside documentFormatFiles are encoded as strings, not integers, and require explicit casting.S-B/A primary document can be far shorter than its corresponding S-B while legally encompassing the same disclosures. Treating document length as a proxy for disclosure completeness will misread amendments.8888 (Foreign Governments) and carry an irsNo of 000000000.cik and order by filedAt.entities array can in principle hold more than one party — for example, a co-issuer or a supplemental filer — though most Schedule B records carry a single filer entity.Form S-B is filed by a narrow class of issuers whose registration content is governed by Schedule B of the Securities Act of 1933 rather than Schedule A. Eligibility is fixed by statute (Securities Act Section 7), not by SEC rule. The filer population is:
The registrant named in the S-B is the sovereign, sub-sovereign, or supranational entity itself. Underwriters and counsel are identified inside the filing but are never the filer.
Form S-B is event-driven, not periodic. A record arises when an eligible issuer initiates or updates a U.S.-registered public offering. Typical triggers:
Each S-B and S-B/A receives its own EDGAR accession number and is a separate record in the dataset.
The disclosure obligation flows directly from the Securities Act of 1933:
The dataset covers S-B and S-B/A filings from January 2002 forward, reflecting the EDGAR availability of the S-B form code in electronic format. Earlier paper-era Schedule B registrations, which date to the original 1933 enactment, are not included.
Form S-B is a Securities Act registration statement filed under Schedule B by foreign sovereigns, their political subdivisions, and supranational entities. The forms most easily confused with it cluster into four groups: (1) the retired Regulation S-B small-business forms that share the "SB" label, (2) the foreign-government Exchange Act forms (Form 18 and 18-K), (3) corporate registration forms used by foreign private issuers (F-series) and domestic issuers (S-series), and (4) adjacent regimes touching sovereign or governmental debt (Form T-3, Rule 144A/Reg S, and the Section 3(a)(2) municipal exemption).
The most important boundary to draw, because the name collision is misleading. SB-1, SB-2, 10-KSB, and 10-QSB were issued under Regulation S-B, a scaled-disclosure regime for small domestic operating companies (originally under $25M in revenues and float). Regulation S-B was rescinded in 2008; smaller reporting companies now use Regulation S-K with scaled accommodations.
Concrete differences:
Any "SB" dataset that mixes these families merges two unrelated regulatory regimes. This dataset covers only Schedule B sovereign registrations.
The closest natural counterparts because they share Form S-B's filer population.
The relationship is sequential, not substitutive. Form S-B is the primary-market Securities Act registration of an offering; Form 18 registers the resulting securities for exchange listing; Form 18-K supplies ongoing fiscal, debt-service, and political disclosure. Form S-B is event-driven (one filing per offering plus S-B/A amendments); 18-K is periodic. A complete sovereign-debt research file requires both.
The F-series registers Securities Act offerings by foreign corporate issuers (FPIs), not by sovereigns. F-1 is the long-form, F-3 the seasoned shelf, F-4 the business-combination form, and F-7/F-8/F-9/F-10/F-80 are the Multijurisdictional Disclosure System forms used primarily by large Canadian issuers.
Overlap with Form S-B is limited to "Securities Act registration by a non-U.S. entity." Disclosure architecture diverges sharply: F-series filings demand Regulation S-K-style corporate content (business description, risk factors, MD&A, executive compensation, audited GAAP/IFRS financials with reconciliation where required). Schedule B has no corporate analog; the issuer is a government, so the disclosure is fiscal accounts, funded-debt schedules, and counsel's opinion on sovereign authority. A foreign private issuer state-owned enterprise with separate juridical personality files on the F-series; the sovereign itself files on S-B. The two implement different statutory schedules (A vs. B) and cannot substitute for each other.
The naming similarity with Form S-B is purely alphabetical. The S-series registers Securities Act offerings by domestic issuers under Schedule A and Regulation S-K. None is available to a foreign government, foreign political subdivision, or supranational; conversely, S-B is unavailable to any domestic corporate issuer. Filer populations and disclosure schedules do not overlap; the only common element is the Securities Act registration trigger itself. (Form S-1, Form S-3, Form S-4, and S-11 are the principal entries in the S-series.)
Form T-3 qualifies an indenture under the Trust Indenture Act of 1939 for debt securities offered without Securities Act registration (typically Section 3(a)(9) or 3(a)(10) exchanges). It is an indenture-qualification filing, not a registration statement, and contains no sovereign fiscal disclosure.
A sovereign issuing registered debt on Form S-B normally files the indenture as an exhibit to the S-B itself rather than on a standalone T-3. T-3 stands alone primarily where the underlying debt is unregistered but still requires a qualified indenture. The two forms touch the same indenture documentation but serve different statutory purposes and are not interchangeable.
A large share of foreign-sovereign USD issuance into the United States is sold under Rule 144A (resales to QIBs) or Regulation S (offshore) and is therefore exempt from Securities Act registration. These offerings produce no Form S-B and no EDGAR registration footprint; their offering memoranda circulate through dealer manuals, Bloomberg, and offshore listing venues such as Euronext Dublin or LuxSE.
The implication for users: Form S-B captures only publicly registered sovereign offerings to U.S. investors. It is authoritative for that perimeter but is not a census of a country's U.S. capital-markets activity, which typically also includes substantial 144A/Reg S volume invisible to EDGAR.
Section 3(a)(2) of the Securities Act exempts securities issued by U.S. states, municipalities, and certain governmental and quasi-governmental issuers. U.S. municipal bonds therefore never file Form S-B or any S-/F-series form; their disclosure flows through MSRB EMMA (official statements and continuing disclosure) rather than EDGAR.
The boundary is strictly jurisdictional: Form S-B is for foreign sovereigns and sub-sovereigns plus supranationals; U.S. sub-sovereign debt sits in the EMMA system. There is no EDGAR overlap.
Form S-B is the only EDGAR registration form whose filers are foreign sovereigns, foreign political subdivisions, and supranational entities, and whose disclosure framework is statutory Schedule B rather than Regulation S-K or its FPI analogues. The dataset is therefore the only corpus capturing the full text of registered foreign-sovereign offering documents filed in the United States — prospectus body, funded-debt schedules, fiscal accounts, counsel's opinion, and underwriter identification — together with S-B/A amendments.
It complements rather than replaces:
The label "S-B" must not be conflated with the retired Regulation S-B small-business forms; they share three letters and nothing else.
Schedule B is a narrow corner of EDGAR covering sovereigns, political subdivisions, and supranationals registering securities into the U.S. The user base is correspondingly tight, drawn from sovereign capital markets, cross-border legal practice, EM credit, and regulatory monitoring.
Sell-side and buy-side credit analysts mine the funded debt schedules, prior-year receipts and expenditures, and securities descriptions (coupon, maturity, currency, redemption, sinking fund, guarantees) for issuer fundamentals not cleanly captured by data vendors. Output: internal credit memos justifying position limits and recovery assumptions, and published research notes priced against the issuer's outstanding curve.
PMs running hard-currency, blended, and macro mandates track which sovereigns and supranationals have registered, on what terms, and through which underwriters. Issuer name, offering size, use of proceeds, and underwriter identification drive index inclusion, liquidity, and primary-market participation calls. S-B/A amendments are watched for last-minute changes to deal terms.
Funded debt tables and prior-year fiscal data are reconciled against central bank, multilateral, and ministry-of-finance figures to feed sovereign rating models. The opinion of counsel and securities description are pulled to verify covenants, governing law, sovereign immunity waivers, and pari passu language for issue-level ratings and recovery analysis.
Bankers use the corpus as a precedent library when pitching mandates: deal structure, denominations, listing, selling restrictions, fees, and underwriter disclosures benchmarked against recent transactions. During live deals, S-B/A amendments are monitored for competing sovereign supply that could move pricing windows.
Issuer and underwriter counsel use the filings as a drafting bank for the description of securities, opinion of counsel, fiscal agency arrangements, governing law, immunity waivers, jurisdiction submission, and tax provisions. HTML and TXT copies support clause-level diffing across precedents and inform negotiations over the depth of fiscal disclosure required in a new prospectus.
Broker-dealer, custodian, and asset manager compliance staff verify the U.S. registration status of sovereign and supranational counterparties before clearing, custody, or distribution. They rely on issuer identification fields, accession metadata, original and amendment filing dates, and the underwriter list to support onboarding files and sanctions screening.
Academics use the 2002-present series for panel data on sovereign supply, contract-design diffusion (pari passu, collective action clauses), and disclosure intensity over time. Structured fields anchor event studies; full text supports text-as-data classification of clauses across issuers.
Reporters confirm issuance facts — issuer, offering size, syndicate, amendment timing, use of proceeds — and trace how often a given government taps the U.S. market under Schedule B.
Engineers parse the JSON metadata for accession numbers, filer IDs, and timestamps, then extract issuer, principal, coupon, maturity, and underwriter fields from the HTML and TXT documents. Outputs feed new-issue trackers, primary-market calendars, supply factors in EM rates models, and CUSIP-to-registration reference tables.
Schedule B prospectuses carry disclosure patterns — immunity waivers, fiscal agency clauses, supranational charter references — that general training data covers poorly. The HTML and TXT documents, anchored by JSON metadata, support precedent-search and clause-extraction systems for sovereign and supranational work.
Each group draws on a different slice — funded debt schedules, securities description, opinion of counsel, underwriter identification, or accession metadata — but shares a need for a complete, historically deep record of Schedule B registrations on EDGAR.
The following workflows are grounded in the Schedule B record structure (sovereign filer entities, fileNo+cik amendment chains, funded debt schedules, opinion of counsel, underwriter identification, and the per-month ZIP layout).
Stream the per-month ZIPs, parse metadata.json for formType, accessionNo, filedAt, and the entities[].companyName/cik/fileNo triple, then extract issuer name, aggregate principal, currency, coupon, and maturity from the cover page and Calculation of Registration Fee table of the primary .htm. Output: a sovereign and supranational primary-market issuance tracker keyed on accession number, with one row per registered tranche and an attached link back to linkToFilingDetails.
Index the prospectus body and exhibit set (form of fiscal agency agreement, form of security, opinion of counsel) across all S-B and S-B/A records to power clause-level retrieval over pari passu, collective action, sovereign-immunity-waiver, governing-law, and submission-to-jurisdiction language. Output: a precedent bank that returns ranked clause excerpts with issuer, year, and accession-number provenance for use during drafting and negotiation of a new prospectus.
Pull the funded debt schedule and the receipts-and-expenditures section out of the primary document for each issuer's most recent S-B (and any post-effective S-B/A pricing supplements), and reconcile the figures against ministry-of-finance and IMF Article IV data. Output: a structured exhibit packaged for sovereign rating committees showing debt composition by instrument, currency, and maturity bucket, with deltas versus the prior registered filing for the same cik.
Group records by entities[].fileNo together with cik to reconstruct each registration's full S-B plus S-B/A chain, then watch new filedAt timestamps within an open chain for pricing supplements, revised fee tables, or restated underwriter syndicates. Output: a syndicate-desk alert feed flagging amendments to active sovereign shelves, useful for repricing competing supply and confirming final deal terms before allocation.
Extract the plan-of-distribution / underwriter-identification block from each primary document, normalize bookrunner and co-manager names, and join to the offering size parsed from the Calculation of Registration Fee table. Output: a Schedule-B-only league table of bookrunners on registered foreign-sovereign and supranational offerings, used as a fact base for EM debt newsletters and as evidence in DCM mandate pitches.
Combine the 2002-to-present series across all sovereign CIKs into a longitudinal panel: one row per accession, columns for issuer, filing date, file number, amendment depth, principal amount, currency, governing law, and presence of specific clause families detected from the prospectus text. Output: a research-grade panel suitable for event studies on contract-design diffusion (collective action clauses, pari passu wording) and for disclosure-intensity comparisons across issuers and decades.
Look up a sovereign or supranational counterparty by cik or companyName, confirm an active Schedule B registration via formType, fileNo, and the latest filedAt in the amendment chain, and capture the underwriter list and U.S. authorized representative from the primary document. Output: an onboarding-file artifact for broker-dealer, custodian, and asset-manager compliance teams documenting U.S. registration status and authorized agent for service of process before clearing or distribution.
Dataset Index JSON API: https://api.sec-api.io/datasets/form-sb-files.json
This endpoint returns the dataset's metadata, including its name, description, last update timestamp, earliest sample date (2002-01-01), covered form types (S-B, S-B/A), container format (ZIP), included file types (HTML, TXT, JSON, PDF), and the download URL for the full archive. It also returns the list of all per-month ZIP containers, each with its own download URL, key, size in bytes, record count, and last updated timestamp. This endpoint does not require an API key.
Use this index to monitor which containers were updated in the most recent refresh run and to decide which monthly shards to download incrementally rather than re-downloading the entire dataset.
Example response:
1
{
2
"datasetId": "1f13365b-9ae0-69b4-b8f7-14424877c766",
3
"datasetDownloadUrl": "https://api.sec-api.io/datasets/form-sb-files.zip",
4
"name": "Form S-B Files Dataset",
5
"updatedAt": "2026-04-18T03:01:04.627Z",
6
"earliestSampleDate": "2002-01-01",
7
"totalRecords": 2339,
8
"totalSize": 72154086,
9
"formTypes": ["S-B", "S-B/A"],
10
"containerFormat": "ZIP",
11
"fileTypes": ["HTML", "TXT", "JSON", "PDF"],
12
"containers": [
13
{
14
"downloadUrl": "https://api.sec-api.io/datasets/form-sb-files/2026/2026-03.zip",
15
"key": "2026/2026-03.zip",
16
"size": 13818783,
17
"records": 154,
18
"updatedAt": "2026-03-21T02:51:19.000Z"
19
}
20
]
21
}
Download Entire Dataset: https://api.sec-api.io/datasets/form-sb-files.zip?token=YOUR_API_KEY
Downloads the complete Form S-B Files Dataset as a single ZIP archive containing all per-month container files from January 2002 to the most recent refresh. This endpoint requires an API key passed via the token query parameter.
Download Single Container: https://api.sec-api.io/datasets/form-sb-files/2026/2026-03.zip?token=YOUR_API_KEY
The dataset is sharded into year/month ZIP containers following the path pattern <YYYY>/<YYYY>-<MM>.zip. Each container holds the metadata file and original EDGAR documents for all S-B and S-B/A filings submitted in that month. Use the per-container downloadUrl from the index response to fetch a specific month. This endpoint requires an API key passed via the token query parameter.
The dataset covers EDGAR filings of Form S-B and its amendment Form S-B/A — the registration statement prescribed under Schedule B of the Securities Act of 1933 for foreign sovereigns, their political subdivisions, and supranational entities offering securities to U.S. investors.
Eligibility is fixed by Section 7 of the Securities Act, not by SEC rule. Eligible filers are foreign sovereign governments (for example, the State of Israel or the United Mexican States), political subdivisions of foreign governments (principally Canadian provinces such as Quebec and Ontario), supranational and multilateral financial institutions (such as IBRD/World Bank, the Inter-American Development Bank, the Asian Development Bank, and the European Investment Bank), and agencies or instrumentalities of foreign governments where the obligation is sovereign in character. U.S. domestic issuers and foreign corporate issuers cannot use this form.
The names share three letters and nothing else. SB-1, SB-2, 10-KSB, and 10-QSB were issued under Regulation S-B, a now-rescinded scaled-disclosure regime for small domestic operating companies; Regulation S-B was eliminated in 2008. Form S-B is a separate, still-active form used exclusively by foreign sovereigns, foreign political subdivisions, and supranationals under statutory Schedule B. This dataset contains only Schedule B sovereign registrations and does not mix in any Regulation S-B small-business filings.
The F-series registers Securities Act offerings by foreign corporate issuers (foreign private issuers) and demands Regulation S-K-style corporate disclosure — business description, risk factors, MD&A, audited GAAP or IFRS financials. Schedule B has no corporate analog; because the issuer is a government, the disclosure is fiscal accounts, funded-debt schedules, securities terms, and counsel's opinion on sovereign authority. A foreign state-owned enterprise with separate juridical personality files on the F-series; the sovereign itself files on Form S-B.
One record corresponds to a single Schedule B filing accepted by EDGAR — either an original S-B registration statement or an S-B/A amendment — identified by its 18-digit accession number. Each record is delivered as a folder containing a metadata.json index plus the original-submission documents EDGAR received (the SGML-wrapped HTML primary registration statement and any non-image exhibits or supplements such as additional HTML files, PDF attachments, or .txt artifacts).
S-B and its S-B/A filings?Group records on the SEC fileNo (formatted as 333-XXXXXX) carried in each entity object together with the filer's cik. Pre-effective and post-effective amendments reuse the original registration's file number, and the filer's CIK is invariant across the chain, so this join recovers the complete S-B plus S-B/A history for a registration. Amendment numbering itself (Amendment No. 1, Amendment No. 2, etc.) is not exposed as a separate metadata field — it appears only on the cover page of the primary document.
Three categories are not packaged into the per-accession folder: image files (.gif, .jpg, .png such as sovereign seals or signature scans, which remain enumerated in metadata.json but whose bytes are not extracted); the redundant full-submission TXT envelope that concatenates every document in the SGML submission; and structured-data side-files (dataFiles and seriesAndClassesContractsInformation are uniformly empty, since Schedule B issuers are outside the XBRL tagging regime and outside the investment-company series-and-class regime).
You will not find them here. A large share of foreign-sovereign USD issuance into the United States is sold under Rule 144A (resales to qualified institutional buyers) or Regulation S (offshore) and is exempt from Securities Act registration; those offerings produce no Form S-B and no EDGAR registration footprint. This dataset is authoritative for publicly registered foreign-sovereign offerings into the U.S., but it is not a census of a country's full U.S. capital-markets activity.