The Form 10-12B Files Dataset is a complete, accession-by-accession archive of every Form 10-12B and Form 10-12B/A submission delivered to EDGAR — the general-purpose registration statement that a U.S. domestic issuer files under Section 12(b) of the Securities Exchange Act of 1934 to register a class of securities in connection with listing on a national securities exchange. One record is one EDGAR submission for one accession number, packaged as a folder containing a metadata.json index plus the primary registration statement and every textual exhibit EDGAR received. The filer is the registrant itself, most often a spin-off subsidiary (SpinCo) being separated from a public parent, but also carve-outs, reorganized debtors, holding companies, and private operating companies pursuing a non-S-1 exchange-listing pathway. Coverage begins on January 1, 1994, the start of the EDGAR era for these filings, and runs through the latest monthly refresh. The dataset is distributed as monthly ZIP containers holding TXT, JSON, HTML, and PDF files preserved in EDGAR's original SGML document wrappers.
Programmatically retrieve the full list of dataset archive files, download URLs and dataset metadata.
Dataset Index JSON API
Download the entire dataset as a single archive file.
Download Entire Dataset:
Download a single container file (e.g. monthly archive) from the dataset.
Download Single Container:
The dataset captures the full text-bearing contents of every EDGAR Form 10-12B and Form 10-12B/A submission. Form 10-12B is the SEC's general-purpose Exchange Act registration form prescribed when no other registration form fits the registrant's circumstances; its dominant practical use is registering a class of common stock that is being distributed to the shareholders of a parent company in a corporate spin-off, carve-out, or similar separation transaction. The "12B" suffix marks the form as filed under Section 12(b) (national securities exchange listing), as opposed to Form 10 filed under Section 12(g) (the over-the-counter or threshold-based variant). Form 10-12B/A is the amendment variant — a re-filed registration statement (or portion thereof) that supplements, corrects, or completes earlier disclosures, frequently in response to SEC Staff comment letters during the review cycle.
Each record is the complete textual submission for one accession number — not a single extracted section, not an event row, and not a normalized table. The dataset covers the entire population of 10-12B and 10-12B/A filings on EDGAR from January 1994 through the latest refresh, and 10-12B/A amendments are stored as separate, self-contained records sitting alongside the original 10-12B record under the same registrant CIK and the same SEC file number. Records are grouped within monthly container ZIPs whose top-level folder is the year-month (e.g. 2025-07/); inside that folder, each accession-number subdirectory is one record.
The substantive disclosure burden of a 10-12B mirrors that of a Form 10. It requires the Item-by-Item information set defined under Regulation S-K and Regulation S-X, including a description of the registrant's business and properties, risk factors, the securities being registered, management and executive compensation, security ownership, certain relationships and related transactions, legal proceedings, market for the registrant's common equity, audited historical financial statements (and, in spin-off contexts, unaudited pro forma financial statements), management's discussion and analysis, changes in and disagreements with accountants, indemnification of directors and officers, and a list of exhibits and financial-statement schedules. In spin-off filings, the bulk of the readable disclosure is typically presented as an "Information Statement" attached as Exhibit 99.1 and incorporated into the Form 10 body by reference.
One record in the Form 10-12B Files Dataset is a single EDGAR submission of either a Form 10-12B or a Form 10-12B/A, captured in full as the registrant transmitted it to EDGAR. Each record materializes as a folder named after the 18-digit EDGAR accession number with dashes stripped (for example, 000175392625001147). The folder contains a metadata.json index produced by sec-api.io plus the textual documents EDGAR received as part of the submission: the primary registration statement, every exhibit document, and any cover-letter or correspondence pieces. Image attachments referenced from the submission (.jpg, .gif, .png files marked with EDGAR type: GRAPHIC) are deliberately omitted from the on-disk record even though they remain enumerated in metadata.json.
At the storage level, every record is a small directory with a fixed shape:
metadata.json describing the submission as a whole and indexing every document EDGAR received;10-12b or 10-12ba token (e.g. g084891_10-12ba.htm);ex<exhibit-number> tokens that mirror EDGAR's exhibit numbering (ex2-1, ex3-1, ex10-1, ex21-1, ex99-1, etc.);The same on-disk shape applies regardless of how many filings a given month contains.
Every non-JSON document in a record is preserved exactly as EDGAR received it, including the SGML-style document wrapper that EDGAR places around each piece of a submission. The wrapper precedes the actual HTML or text payload:
1
<DOCUMENT>
2
<TYPE>10-12B/A
3
<SEQUENCE>1
4
<FILENAME>g084891_10-12ba.htm
5
<DESCRIPTION>10-12B/A
6
<TEXT>
7
<HTML>
8
... full HTML body of the registration statement ...
9
</HTML>
10
</TEXT>
11
</DOCUMENT>
The <TYPE>, <SEQUENCE>, <FILENAME>, and <DESCRIPTION> header lines correspond directly to the matching entries in metadata.json -> documentFormatFiles, so the JSON index and the SGML headers can be cross-checked against each other. Exhibits carry the same wrapper with <TYPE>EX-2.1, <TYPE>EX-3.1, <TYPE>EX-10.1, <TYPE>EX-99.1, etc. The <TEXT> block holds the document body verbatim (HTML, ASCII text, or PDF stream); programmatic readers expecting pure HTML must skip the SGML header lines or parse the wrapper explicitly.
The combined "Complete submission text file" (<accession>.txt), which concatenates every document with its SGML wrapper into a single ASCII stream, is referenced from metadata.json -> linkToTxt but is not duplicated as a separate file inside the folder; the per-document files already carry the same content split out.
metadata.json schemaThe metadata.json for one record is a structured, self-contained description of the EDGAR submission. The fields that carry intentional, consumable meaning are:
formType — the submission form code, exactly one of 10-12B or 10-12B/A for this dataset.accessionNo — the dashed 18-digit accession number (e.g. 0001753926-25-001147).description — a short human-readable form description.filedAt — the EDGAR acceptance timestamp with timezone offset.linkToFilingDetails — sec.gov URL to the primary filing document.linkToTxt — sec.gov URL to the combined submission .txt file.linkToHtml — sec.gov URL to the EDGAR *-index.htm page for the submission.linkToXbrl — sec.gov URL to XBRL data when present.id — the sec-api.io internal record identifier (32-character hex).documentFormatFiles — an ordered array enumerating every document EDGAR received in the submission. Each element carries sequence (EDGAR sequence number, with a single space for non-numbered items such as the combined .txt), size (byte size as a string), documentUrl (sec.gov URL), description (EDGAR description such as 10-12B/A, EXHIBIT 99.1, GRAPHIC), and type (EDGAR document type such as 10-12B/A, EX-99.1, EX-3.1, GRAPHIC). The first entry, with sequence: "1", is always the primary form document; subsequent numbered entries are exhibits in EDGAR sequence; an entry with sequence: " " is the combined submission text file.dataFiles — an array of structured-data files (XBRL instance, schema, calculation/definition/label/presentation linkbases). For 10-12B records this array is typically empty.seriesAndClassesContractsInformation — an array used to capture mutual-fund series and class contract information; empty for typical 10-12B filings, which are operating-company spin-off registrations rather than fund filings.entities — an array of one or more parties associated with the submission, almost always a single registrant for 10-12B. Each entity carries companyName (with the role suffix appended in parentheses, e.g. Cortigent, Inc. (Filer)), cik (zero-padded Central Index Key), type (the entity's form-type association, e.g. 10-12B/A), act (Securities Act under which the form is filed, e.g. 34 for the Exchange Act), fileNo (SEC file number, e.g. 001-41716, where the 001- prefix is the National Securities Exchange file-number range used for Section 12(b) registrants), filmNo (SEC film number assigned at acceptance), irsNo (IRS Employer Identification Number), sic (SIC industry code and description, with HTML-entity-encoded ampersands such as &), stateOfIncorporation (two-letter state or country code), fiscalYearEnd (MMDD string such as 1231), and tickers (array of ticker symbols when the registrant already trades).The file types found in the dataset are TXT, JSON, HTML, and PDF. The dominant pattern in modern records is per-document .htm/.html files plus the JSON manifest; .txt documents appear in older or text-only filings, and .pdf files appear only when a registrant supplements the submission with a graphical or scanned exhibit.
The primary document (<TYPE>10-12B or <TYPE>10-12B/A, <SEQUENCE>1) is the registration statement proper. Its body, which lives between the <TEXT> opening tag and </TEXT> closing tag of the SGML wrapper, follows the Item structure prescribed for Form 10:
In spin-off-driven 10-12B filings, the substantive disclosure body is frequently presented as a separate "Information Statement" document attached as Exhibit 99.1 and incorporated into the Form 10 by reference. In that pattern, the primary 10-12B document is short — cover page, item-by-item cross-references that point into the Information Statement, the exhibit index, and signatures — while the bulk of the readable narrative (business description, risk factors, MD&A, financial statements, descriptions of capital stock, and the mechanics of the separation transaction) lives in the ex99-1 document.
Exhibits are stored as separate per-document files in the same folder, each wrapped in its own <DOCUMENT> SGML block with a <TYPE>EX-<n> header. The exhibit index in Item 15 of the primary document lists each exhibit by number and description. The exhibits typically present in a 10-12B include:
Each exhibit document carries the EDGAR <TYPE> and <DESCRIPTION> headers (e.g. <TYPE>EX-99.1, <DESCRIPTION>EXHIBIT 99.1) that are mirrored verbatim in the corresponding documentFormatFiles[] entry in metadata.json.
Each record includes: the metadata.json index; the primary registration statement document; every exhibit document EDGAR received in textual form; any correspondence or cover-letter documents; and the SGML wrappers that bind type/sequence/filename/description metadata to each document body.
Each record excludes: the combined <accession>.txt "Complete submission text file" (referenced via linkToTxt rather than duplicated, since the per-document files already split the same content out); and image attachments such as .jpg, .gif, and .png files (which remain listed in metadata.json -> documentFormatFiles with type: GRAPHIC and description: GRAPHIC but are not stored on disk). External materials that EDGAR rules permit a registrant to incorporate by reference — for example, audited financial statements previously filed by a parent company — are not pulled into the record either; only the documents physically transmitted in the 10-12B accession are present.
Form 10-12B/A submissions appear as their own records, each with its own accession number, its own folder, and its own metadata.json whose formType is 10-12B/A. They are not differentially encoded against the original 10-12B; they are full re-filings of the registration statement (or a defined portion of it) with whatever changes the registrant has made, again wrapped in EDGAR's standard SGML document blocks. The chain of an original 10-12B plus one or more 10-12B/A records can be reconstructed by joining on the registrant CIK (in entities[].cik) and the SEC file number (in entities[].fileNo), since EDGAR assigns the same 001-xxxxx file number to every amendment in the same registration thread, and ordering by filedAt. Successive amendments tend to grow shorter than the original as registrants address Staff comments narrowly, but pre-effective amendments commonly re-file the entire Information Statement (ex99-1) with marked-up changes.
The Form 10 / 10-12B disclosure framework has been substantially enriched since EDGAR began accepting filings in 1994. Three regulatory changes are most visible across the dataset's coverage:
Spin-off-specific disclosures — particularly the unaudited pro forma financial statements giving effect to the separation, capitalization, and distribution — have been a stable element of 10-12B filings throughout the dataset's coverage, though their presentation has evolved with successive updates to Article 11 of Regulation S-X. The 2020 amendments under Release 33-10786 restructured the pro forma framework into "Transaction Accounting Adjustments," "Autonomous Entity Adjustments," and optional "Management's Adjustments," and recent records reflect that revised structure.
Form 10-12B filings span EDGAR's full document-format history:
.txt documents wrapped in the same <DOCUMENT>...</DOCUMENT> SGML blocks. Tables are rendered as fixed-width ASCII, and graphics referenced from the document are absent from the dataset by design..htm files with the SGML wrapper preserved on top. PDF documents appear occasionally for graphical or scanned exhibits.metadata.json -> linkToXbrl and dataFiles are typically empty for 10-12B records, and structured financial-data tagging is generally not present even on recent filings. When XBRL data does accompany a 10-12B — for instance because the financial statements were already tagged in a related filing — it appears as additional entries in dataFiles and as <TYPE>EX-101.* documents in the SGML stream.A few practical points matter when extracting or analyzing these records:
<HTML> (or <html>) tag and ignore everything after the closing </HTML>; alternatively, they can parse the <DOCUMENT> block to recover <TYPE>, <SEQUENCE>, <FILENAME>, and <DESCRIPTION> headers without consulting metadata.json.metadata.json -> documentFormatFiles is a complete index of what EDGAR received, but image entries (type: GRAPHIC) will not have a corresponding file on disk; consumers should resolve those via documentUrl to sec.gov when needed.documentFormatFiles[] entry (sequence: "1") is always the primary 10-12B or 10-12B/A document. In spin-off filings, however, the substantive disclosure narrative typically lives in EX-99.1, not in the primary document. Consumers extracting "the registration statement text" should follow the cross-reference pattern rather than treating the primary document alone as the disclosure body.entities[].sic are HTML-entity-encoded (& rather than &); downstream processing should decode them.001- indicate Section 12(b) registrants destined for a national securities exchange listing, which is consistent with the 10-12B form's purpose; this can be used as a sanity check on the entity record.formType (and the matching <TYPE> header in the SGML wrapper). Reconstructing an amendment chain requires joining on CIK and SEC file number across records and ordering by filedAt.metadata.json -> accessionNo carries the dashed canonical form. Joining records to other EDGAR datasets requires reconciling those two representations.The filer is the issuer (registrant) itself — a company registering a class of its equity securities under Section 12(b) of the Securities Exchange Act of 1934 in connection with listing that class on a national securities exchange. The registration statement is signed by the registrant and by an authorized officer and a majority of its board, as required by Form 10's signature instructions and Rule 12b-11.
Typical filer profiles:
Persons named in the filing — the parent company, directors and officers, principal shareholders — are subjects of disclosure, not filers. The parent in a spin-off makes its own disclosures (typically a Form 8-K and an information statement to shareholders) but does not sign the SpinCo's Form 10-12B.
Foreign private issuers generally register on Form 20-F instead and are rare in this dataset. Investment companies use the N-series forms and do not file Form 10-12B.
These records are filed (not furnished) and are subject to Section 18 liability and to SEC staff review and comment.
Form 10-12B is transaction-driven, not periodic. The trigger is the registrant's intent to have a class of its equity listed on a national securities exchange in a context where the listing — rather than a cash offering — is the operative event:
The Form 10-12B is the legal instrument that effects Section 12(b) registration; the underlying business event is what makes it necessary at that time.
The dataset's earliest records date to January 1994, reflecting EDGAR rollout. Section 12(b) registrations predate EDGAR, but earlier filings exist only on paper.
A 10-12B/A is an amendment to a previously filed 10-12B, linked by SEC file number. Amendments are filed to:
A single 10-12B commonly goes through several /A amendments before the registrant requests acceleration. The dataset contains the original and each amendment as distinct accession-numbered records.
Form 10-12B registers a class of securities under Section 12(b) of the Securities Exchange Act of 1934 in connection with listing on a national securities exchange. The forms most often confused with it cluster in four groups: the other Exchange Act registration form (10-12G), the Securities Act offering registrations (S-1, S-4), post-registration periodic and current reports (10-K, 8-K), shareholder-communication filings (Schedule 14C, DEF 14A), and the foreign-issuer counterpart (20-F).
Form 10-12B is narrowly the Section 12(b) class-registration document for U.S. domestic issuers listing on a national exchange, used most often for spin-offs. It is not an offering registration (S-1, S-4), not a periodic or current report (10-K, 8-K), not a shareholder communication (14C, DEF 14A), and not the foreign-issuer equivalent (20-F). Its only true twin is 10-12G, which differs solely on listing venue. A complete picture of a spin-off typically requires the parent's 8-K, any 14C or DEF 14A, the new entity's 10-12B, and its subsequent 10-K — each capturing a different slice that 10-12B alone does not provide.
Form 10-12B is the first comprehensive disclosure document filed by a company becoming a public reporting issuer through a spin-off or carve-out. Each user group below reads a different slice of the filing.
Coverage and structuring teams on separation, split-off, and Reverse Morris Trust mandates use the dataset as a precedent library. They filter formType and filedAt, build peer sets by industry and size from entities, then pull "Reasons for the Separation," capitalization, pro forma financials, and the separation, tax matters, and transition services agreement exhibits to benchmark SpinCo leverage, dividend policy, TSA pricing and duration, and indemnification caps. Output: fairness-opinion support, structuring memos, and pitch decks anchored to comparable deals.
Outside counsel and in-house securities lawyers mine the corpus for risk factor language on Section 355 qualification, anti-Morris-Trust restrictions, and successor liability, and pull the full exhibit set — separation, tax matters, employee matters, TSA, IP cross-license, registration rights — to compare drafting conventions for indemnity baskets, survival periods, and step-in rights. Output: working-draft markup and negotiation positions.
Sell-side and buy-side fundamental analysts use the 10-12B as the only complete financial document available before "when-issued" trading and the first standalone 10-K. They extract carve-out historical financials, pro forma adjustments, segment disclosure, and standalone-cost MD&A to seed initiation models, and use "Description of Capital Stock" and "Description of Indebtedness" for share count, debt schedule, and interest expense. Output: initiation report, target price, and call timed to distribution.
Spin-off arbitrage and special-situations PMs track amendment chains (10-12B/A) to monitor SEC review, screen risk factors and "Reasons for the Separation" for parent index-eligibility forced selling and ParentCo deleveraging mechanics, and read the tax matters agreement for the two-year acquisition and buyback restrictions that constrain SpinCo capital action. Output: position thesis sized to the distribution calendar.
Sponsors underwriting corporate carve-outs use the employee matters agreement to model pension and benefit transfers, the TSA to size stranded-cost emergence, and the IP cross-license to flag shared assets that complicate financing. Pro forma financials and standalone-cost MD&A are used to pressure-test management projections in diligence. Output: bid models, separation cost estimates, and Day-1 / Day-100 plans.
In-house strategy and corp dev teams pull peer 10-12Bs to benchmark dis-synergy disclosure, SpinCo board composition, charter and bylaw anti-takeover provisions, and retained-stake mechanics. The "Relationship with Parent" section and intercompany agreement schedule scope the internal workstreams. Output: board paper recommending spin, sale, IPO, or deferral.
Researchers studying registration practice, separation governance, and tax-free reorganizations use the full 1994-to-present history with entities, filedAt, and amendment chains to build longitudinal samples — quantifying staff comment cycles, classifying risk factors over time, and tracking tax matters agreement language across IRS guidance changes. Output: empirical papers and policy commentary.
IR and reporting teams at issuers approaching their own separation benchmark "Why the Separation" framing, segment renaming, and pro forma presentation against recent peer filings, and pull the exhibit set to anticipate analyst-day questions on TSA exit, dis-synergies, and capital allocation. Output: SpinCo investor deck, Form 10 narrative, and Q&A briefing books.
Credit and forensic professionals focus on the carve-out basis-of-presentation footnote, corporate expense allocations, pension and OPEB transfers, and indemnification and contingent liability schedules. Pro forma capitalization, debt schedules, and covenant disclosures feed standalone ratings and bond-pricing models. Output: credit opinions, indenture commentary, and forensic red flags.
Document-intelligence and quant research teams use the corpus as training and evaluation data: a controlled population of registration filings with a stable section taxonomy (Items 1, 1A, 2, 4, 5, 6, 13, 15) across three decades, plus formType, filedAt, and entities for stratified sampling. Workloads include long-context fine-tuning, RAG over registration prose, clause classifiers for tax matters and TSA terms, and table-extraction benchmarks against pro forma statements and exhibit indices.
The following workflows reflect how practitioners actually use the Form 10-12B Files Dataset. Each one ties to specific record components (metadata.json fields, primary registration document Items, and named exhibits) and produces a concrete output.
Pull every EX-2.1 (separation and distribution agreement) and the EX-10 series carrying transition services, tax matters, employee matters, and IP cross-license agreements across all 10-12B records, joined to registrant industry via entities[].sic and deal date via filedAt. Compare TSA fee mark-ups, service-tail durations, indemnity baskets and caps, and tax-matters two-year acquisition/buyback restrictions across recent spin-offs. Output: a clause-level precedent library with peer-set comparables that disclosure counsel and bankers cite during drafting and structuring negotiations.
Reconstruct each registration thread by joining records on entities[].cik plus entities[].fileNo (the 001- Section 12(b) file number), ordering by filedAt, and counting successive formType: 10-12B/A filings. Diff each amendment's EX-99.1 Information Statement against the prior version to localize Staff-driven changes (risk factors added, MD&A reworked, pro forma adjustments restated). Output: a Staff-comment-cycle dataset for measuring review duration, identifying SpinCos with stale financials triggering re-filings, and timing event-driven positions to expected effectiveness dates.
For each spin-off record, parse Item 13 of the primary document and the financial-statements section of EX-99.1 to capture unaudited pro forma condensed balance sheets, capitalization tables, debt schedules, and standalone-cost MD&A under Article 11 of Regulation S-X (including the post-2020 split into Transaction Accounting Adjustments, Autonomous Entity Adjustments, and Management's Adjustments). Combine with EX-21 for the post-separation legal-entity footprint. Output: pre-distribution initiation models, share-count and interest-expense schedules, and credit-rating inputs available before the SpinCo's first 10-K.
Use formType, filedAt, and entities[].sic to stratify samples across the 1994-to-present coverage, then split each record into Item 1 (Business), Item 1A (Risk Factors), Item 2 (MD&A), Item 6 (Executive Compensation), and the EX-99.1 Information Statement using the SGML <TYPE>/<DESCRIPTION> headers. Train classifiers for tax-matters and TSA clause types, risk-factor topic taxonomies (Section 355, anti-Morris-Trust, successor liability), and table extractors against pro forma and exhibit-index structures. Output: fine-tuned models, RAG indexes, and evaluation benchmarks for registration-document NLP.
For an issuer preparing its own spin-off, filter peer 10-12B records by entities[].sic and recent filedAt, then extract the "Reasons for the Separation," "Relationship with Parent," dis-synergy quantification, and Item 1A risk factors from each peer's EX-99.1. Compare segment renaming, capital-allocation framing, and TSA-exit narrative. Output: SpinCo investor deck language, Form 10 narrative drafting, and analyst-day Q&A briefing books anchored to recent precedent.
Aggregate EX-21 (subsidiaries) lists, EX-3.1/EX-3.2 (charter and bylaws as effective on registration), and Item 4 beneficial-ownership tables to chart post-separation legal entity trees, anti-takeover provisions (classified board, supermajority, poison pill mechanics), and retained-parent-stake structures. Cross-reference with entities[].stateOfIncorporation to study Delaware-versus-other governance choices. Output: governance benchmarking for corp dev board papers and academic studies of separation-era charter design.
Read the basis-of-presentation footnote and corporate-expense-allocation disclosures inside the audited carve-out financials in EX-99.1, paired with pension/OPEB transfer mechanics in the employee matters agreement (EX-10) and indemnification schedules in the separation agreement (EX-2.1). Combine with pro forma debt schedules and covenant disclosures from Item 13. Output: standalone credit opinions, indenture commentary, and forensic red-flag lists for SpinCo bond pricing and post-distribution surveillance.
The dataset is accessible through three complementary endpoints. The index endpoint is open and returns dataset metadata; the archive and per-container endpoints require an API key.
Dataset Index JSON API: https://api.sec-api.io/datasets/form-1012b-files.json
Returns dataset-level metadata (name, description, earliest sample date, form types covered, container format, file types, last update timestamp, total records, total size) along with the full list of container files. Each container entry includes its download URL, key, size in bytes, record count, and last updated timestamp. Use this endpoint to monitor which containers were refreshed in the latest run and to decide which files to download incrementally. This endpoint does not require an API key.
Example response:
1
{
2
"datasetId": "1f13365b-9ae0-697b-b59b-2217c1497c89",
3
"datasetDownloadUrl": "https://api.sec-api.io/datasets/form-1012b-files.zip",
4
"name": "Form 10-12B Files Dataset",
5
"updatedAt": "2026-04-25T02:56:10.779Z",
6
"earliestSampleDate": "1994-01-01",
7
"totalRecords": 17740,
8
"totalSize": 801534093,
9
"formTypes": ["10-12B", "10-12B/A"],
10
"containerFormat": "ZIP",
11
"fileTypes": ["TXT", "JSON", "HTML", "PDF"],
12
"containers": [
13
{
14
"downloadUrl": "https://api.sec-api.io/datasets/form-1012b-files/2025/2025-07.zip",
15
"key": "2025/2025-07.zip",
16
"size": 13818783,
17
"records": 154,
18
"updatedAt": "2026-04-25T02:56:10.779Z"
19
}
20
]
21
}
Download Entire Dataset: https://api.sec-api.io/datasets/form-1012b-files.zip?token=YOUR_API_KEY
Downloads the complete dataset as a single ZIP archive containing every Form 10-12B and 10-12B/A filing from January 1994 to the latest refresh. This endpoint requires an API key.
Download Single Container: https://api.sec-api.io/datasets/form-1012b-files/2025/2025-07.zip?token=YOUR_API_KEY
Containers are organized by year and month (e.g. 2025/2025-07.zip) and can be downloaded individually instead of pulling the full archive. Container paths are discoverable through the index JSON API. This endpoint requires an API key.
The dataset covers exactly two EDGAR form types: 10-12B, the general-purpose Exchange Act registration statement filed under Section 12(b) of the Securities Exchange Act of 1934 to register a class of securities for listing on a national securities exchange, and 10-12B/A, the amendment variant of that registration statement.
One record is a single EDGAR submission for one accession number, materialized as a folder named after the 18-digit accession number with dashes stripped. The folder contains a metadata.json index plus the textual documents EDGAR received — the primary registration statement, every exhibit document (e.g. EX-2.1 separation agreements, EX-10 material contracts, EX-99.1 Information Statements), and any correspondence — each preserved in its original EDGAR SGML document wrapper.
The filer is the issuer (registrant) itself, signed by an authorized officer and a majority of its board. The dominant population is spin-off subsidiaries (SpinCos) being separated from a public parent through a pro rata distribution; other filers include carve-out and tracking-stock entities, reorganized debtors emerging from Chapter 11, holding companies formed in a reorganization, and private operating companies pursuing a non-S-1 exchange-listing pathway.
The dataset's earliest records date to January 1, 1994, reflecting EDGAR rollout for these filings, and coverage extends through the latest monthly refresh. Section 12(b) registrations predate EDGAR, but earlier filings exist only on paper and are not part of this dataset.
The dataset is distributed as monthly ZIP containers organized by year and month (e.g. 2025/2025-07.zip). Inside each container, every accession-number subdirectory holds a metadata.json plus the submission's documents in TXT, HTML, or PDF form, all wrapped in EDGAR's original <DOCUMENT>...</DOCUMENT> SGML headers. Image attachments (.jpg, .gif, .png) are intentionally excluded from disk but remain enumerated in metadata.json -> documentFormatFiles.
Both are Exchange Act class-registration forms, but 10-12B is filed under Section 12(b) when the class will be listed on a national securities exchange (NYSE, Nasdaq, etc.), while 10-12G is filed under Section 12(g) for issuers crossing asset/holder thresholds without an exchange listing (typically OTC). Disclosure content is essentially identical; only the listing venue determines which form applies. 10-12B effectiveness is governed by Section 12(d) and exchange certification; 10-12G effectiveness runs 60 days after filing.
A 10-12B/A is an amendment linked to its original 10-12B by SEC file number. To reconstruct an amendment chain, join records on entities[].cik (the registrant Central Index Key) and entities[].fileNo (the 001-xxxxx Section 12(b) file number, which EDGAR assigns to every amendment in the same registration thread) and order by filedAt. Each amendment is stored as a separate, self-contained record with its own accession number, not as a diff against the original.