The Form F-80 Files Dataset is a complete corpus of EDGAR submissions of Form F-80 and Form F-80/A — Securities Act registration statements filed under the U.S.–Canada Multijurisdictional Disclosure System (MJDS) by eligible Canadian foreign private issuers to register securities offered as consideration in cross-border exchange offers and business combinations. Each record is a single EDGAR accession packaged as a per-accession folder containing a metadata.json header file plus the original submission documents (HTML, TXT) reproducing the filing as transmitted to the Commission. Filers are MJDS "substantial issuer" Canadian registrants meeting the C$75 million public-float and less-than-40-percent U.S.-ownership thresholds; subjects, target shareholders, and U.S. holders described in the document are not filing parties. Coverage begins January 1, 2002 — the practical start of mandatory EDGAR filing for foreign private issuers — and the dataset is distributed as monthly ZIP containers with TXT, JSON, and HTML file types.
Programmatically retrieve the full list of dataset archive files, download URLs and dataset metadata.
Dataset Index JSON API
Download the entire dataset as a single archive file.
Download Entire Dataset:
Download a single container file (e.g. monthly archive) from the dataset.
Download Single Container:
The dataset packages every EDGAR submission of Form F-80 (initial registration statement) and Form F-80/A (amendment) as an independently retrievable record. Form F-80 is a registration statement under the Securities Act of 1933 used by eligible Canadian foreign private issuers to register securities offered in exchange offers or business combinations under the Multijurisdictional Disclosure System. Eligibility presumes incorporation or organization in a Canadian jurisdiction, a public float of at least C$75 million, and less than 40 percent U.S. ownership of the class of securities subject to the offer. Its distinguishing feature is that the substantive disclosure document is the Canadian offering or take-over bid circular prepared under provincial securities law, filed in the United States substantially as prepared for Canadian regulators and shareholders, without U.S. GAAP reconciliation of the financial statements. Form F-80/A is the amendment variant, used to add, supplement, or revise content already on file under an existing F-80 file number.
The filing is a hybrid object: an EDGAR-formatted U.S. registration statement on the outside (with a U.S. cover page, signatures of the registrant and its U.S. agent for service, fee disclosures, and Securities Act exhibit conventions) wrapped around a Canadian disclosure document on the inside (the take-over bid circular, exchange offer prospectus, information circular, or business combination proxy/management information circular as prepared under the relevant provincial regime). The dataset preserves both layers as filed.
Coverage begins January 1, 2002, which is the practical start of electronic coverage because mandatory EDGAR filing for foreign private issuers, including MJDS filers, took effect in November 2002. Earlier F-80 filings exist only in paper form in SEC historical files and are not included. The dataset is distributed as a ZIP container archive partitioned by month, with TXT, JSON, and HTML file types inside.
One record is a single EDGAR submission of a Form F-80 or Form F-80/A registration statement, packaged as a per-accession folder. The folder is named with the 18-digit, dash-stripped accession number (for example, 000104746913007811) and contains exactly one metadata.json describing the EDGAR header plus the original submission documents reproducing the filing as transmitted to the Commission. The record unit is the filing itself — not the filer, not an individual exhibit, and not an extracted disclosure section. An amendment to a previously filed F-80 produces an additional, independently retrievable record under its own accession with form type F-80/A.
Records are partitioned by year and then by month. Each monthly ZIP unpacks to a directory of the form YYYY/YYYY-MM/<accession>/, where <accession> is the dash-stripped 18-digit EDGAR accession number. Inside that accession folder a record consists of:
metadata.json describing the EDGAR submission header and the document inventory..txt payloads on earlier ones. Each non-metadata file retains the SGML <DOCUMENT> envelope from the original submission.Image attachments referenced by the submission (JPEG, GIF) are intentionally omitted from the dataset; their identity, size, and ordinal position remain catalogued inside metadata.json even though the binaries themselves are not on disk.
The metadata file is a flat JSON object that flattens the EDGAR submission header into structured fields. The principal members are:
formType — "F-80" for an initial registration statement or "F-80/A" for an amendment.accessionNo — the canonical dashed accession (e.g., "0001047469-13-007811").filedAt — ISO 8601 timestamp with timezone offset capturing the EDGAR acceptance moment.periodOfReport — the reporting date associated with the filing in YYYY-MM-DD form.description — the EDGAR human-readable description, which carries the [Amend] marker for amendments.linkToFilingDetails, linkToTxt, linkToHtml, linkToXbrl — canonical SEC.gov URLs to the primary document, the complete-submission TXT package, the EDGAR filing index page, and the XBRL package. linkToXbrl is consistently an empty string for F-80, since the form carries no XBRL exhibits.id — an internal 32-character hex identifier for the record.documentFormatFiles[] — an ordered array enumerating every attachment in the original EDGAR submission. Each element carries sequence (ordinal position), size (byte size as a string), documentUrl (SEC.gov URL), description (short label), and type (EDGAR document type such as F-80/A, EX-5.1, EX-23.1, GRAPHIC). The terminal entry of this array represents the complete-submission TXT package and uses a single-space " " for both sequence and type.entities[] — an array of filer/subject entities. Each entity carries companyName with a role suffix in parentheses (e.g., "... (Filer)", "... (Subject)"), cik, type (the entity's role-specific submission type), sic with both code and label, stateOfIncorporation using EDGAR's state/country codes (Canadian provinces appear as A0–A9 and B0, e.g., A6 for Ontario), act ("33" for the Securities Act of 1933), fileNo (the assigned 333-… file number), filmNo, irsNo (often "000000000" for non-U.S. filers without an IRS EIN), and fiscalYearEnd in MMDD form.seriesAndClassesContractsInformation[] — present in the schema but typically empty for F-80; reserved for investment-company series/class identifiers that this form does not require.dataFiles[] — present in the schema but typically empty for F-80; reserved for XBRL and other structured data attachments that this form does not require.documentFormatFiles[] and entities[] are the load-bearing structures for navigating a record. The first enumerates every constituent file — including image files that are referenced but not packaged on disk — so it is the authoritative inventory of the original submission. The second enumerates the registrant and any co-filers, subject company, or affiliated parties with their EDGAR-assigned identifiers.
Each non-metadata file inside the accession folder preserves the original EDGAR SGML <DOCUMENT> envelope. The envelope opens with <DOCUMENT> and a small header block of <TYPE>, <SEQUENCE>, <FILENAME>, and <DESCRIPTION> lines whose values mirror the matching entry in documentFormatFiles[]. The body is delimited by <TEXT> and </TEXT> and contains either an HTML payload (<HTML> with <HEAD> and <BODY>) or, for older or simpler attachments, plain ASCII text. The closing </DOCUMENT> tag terminates the wrapper. These SGML headers are authoritative for identifying exhibit roles and should be parsed in tandem with the corresponding documentFormatFiles[] entry.
The HTML payload of the primary document is the registration statement itself, structured in the order a reader encounters it:
<DOCUMENT> files within the same accession or be embedded within the primary HTML.Each separate exhibit attachment in the EDGAR submission becomes its own SGML-wrapped file inside the accession folder, with <TYPE> reflecting the EDGAR exhibit type code (for instance EX-5.1, EX-23.1, EX-99.1) and <SEQUENCE> matching the ordinal in documentFormatFiles[].
Each record includes the full EDGAR submission header in structured form (metadata.json) and the complete textual and HTML content of the registration statement and its attached exhibits as originally filed, preserved with their SGML document wrappers. Both initial F-80 filings and F-80/A amendments are included, each as its own record under its own accession. Where a registrant has filed multiple amendments under the same 333-… file number, all of those filings appear as separate, independently retrievable records.
Image binaries listed in the EDGAR submission as GRAPHIC documents (typically .jpg and .gif files used for logos, signature scans, charts, and embedded graphics within the offering document) are not packaged in the dataset, even though the offering document HTML may reference them. Their identities, sizes, and ordinal positions remain in documentFormatFiles[]. Records do not include post-effectiveness materials filed under different form types (such as Form F-X powers of attorney filed separately, prospectuses filed under Rule 424, or post-effective amendments filed on other forms), nor do they include EDGAR correspondence (UPLOAD/CORRESP) or staff comment letters that may relate to the registration statement.
Form F-80 has been a stable form since its adoption alongside the rest of the MJDS in 1991, and the dataset's coverage begins in January 2002. The most consequential change to record content within that coverage is the migration of Canadian financial reporting to IFRS as adopted in Canada effective January 1, 2011 for most publicly accountable enterprises; from that point onward, F-80 records typically embed IFRS financial statements rather than Canadian GAAP statements, while the no-reconciliation principle of MJDS continues to apply. Cover-page conventions have evolved with general SEC amendments to fee-table presentation, but Form F-80 itself remains outside the iXBRL fee-tagging regime, so fee tables continue to appear as static cover-page content rather than tagged Exhibit 107 data. Disclosure expectations regarding cybersecurity, sanctions, and similar topical risk areas occasionally surface in the offering document body when the underlying provincial requirements or transactional context call for them, but no F-80-specific item additions have reshaped the record's structure.
Records from the early years of the dataset (2002 through roughly the mid-2000s) more frequently consist of plain-text submissions with ASCII-rendered tables, whereas later records are essentially uniformly HTML payloads with embedded styling and inline imagery references. The SGML <DOCUMENT> envelope has remained constant across the entire coverage period, so the per-file structure is stable even as the inner payload moved from text to HTML. Form F-80 has never required XBRL or inline XBRL data, so no dataFiles[] entries or XBRL companion files appear in any era of the dataset, and linkToXbrl is consistently an empty string.
Several characteristics of F-80 records affect how they should be read and processed:
F-80/A may either restate the entire offering document or supply only changed pages, exhibits, or supplements. Reconstructing the operative disclosure as of any given date may therefore require reading the initial F-80 together with each subsequent F-80/A under the same 333-… file number.entities[] frequently contains more than one party — a filer, a subject company, and sometimes co-registrants — and the role suffix in companyName is the most reliable indicator of which party is which.documentFormatFiles[] is the canonical inventory: presence of an entry there does not guarantee presence of the file on disk (GRAPHIC rows are catalogued but excluded), so consumers should reconcile the inventory against the filesystem when iterating records.The filer is always the Canadian issuer offering its own securities as consideration. Targets, target shareholders, and U.S. holders described in the document are not filing parties.
Form F-80 is restricted to a narrow class of Canadian foreign private issuers eligible to use the Multijurisdictional Disclosure System as a "substantial issuer." To file, the registrant must satisfy all of the following General Instruction tests:
Canadian issuers that fail any criterion (recently listed entities, smaller floats, high U.S. ownership, certain crown corporations or schedule banks) cannot use Form F-80 and must use Form F-4, Form S-4, or an exemption.
The signing population is the registrant itself, signing officers and directors required for a Securities Act registration statement, and the U.S. authorized representative for service of process. Counsel, auditors, and other experts whose consents appear as exhibits are not registrants.
Form F-80 is event-driven, not periodic. A filing is triggered when an MJDS-eligible Canadian issuer extends into the United States either:
The Section 5 trigger is the dissemination of the Canadian takeover bid circular, exchange offer document, or proxy/information circular to U.S. holders. The initial Form F-80 must be on file no later than the date those documents are first sent or given to U.S. holders. Form F-80 substitutes the Canadian-prepared offering or proxy document for the U.S.-style prospectus that an F-4 or S-4 would require.
A Form F-X (appointment of agent for service of process) must be on file before the F-80 is usable. Form F-X is a separate EDGAR submission and is not part of the F-80 accession.
Form F-80 was adopted in SEC Release No. 33-6902 (June 21, 1991) as part of the MJDS. Mandatory EDGAR filing for foreign private issuers, including MJDS filers, took effect in November 2002, which sets the practical start of electronic coverage for this dataset. Earlier F-80 filings exist only in paper form in SEC historical files.
Form F-80 sits inside the Multijurisdictional Disclosure System, which lets eligible Canadian issuers satisfy U.S. registration and tender-offer rules largely through Canadian disclosure documents. The most useful comparisons are the other MJDS forms (F-7, F-8, F-9, F-10, F-X, Schedule 14D-1F, Schedule 13E-4F) and the non-MJDS analogues for the same transactions (F-4 and S-4). Boundaries between them turn on transaction type, eligibility tier, filer domicile, and U.S. ownership of the subject class.
The closest sibling. F-8 and F-80 register the same transactions (exchange offers, business combinations) for eligible Canadian issuers under MJDS, both relying on Canadian offering and proxy documents. The split is the eligibility tier: F-8 is for issuers meeting the higher public float test, while F-80 covers the lower tier (historically C$75 million public float, less than 40% U.S. ownership of the subject class). A given transaction is filed on one or the other, never both.
Also an MJDS Securities Act form, but limited to rights offerings to existing security holders. Form F-7 does not cover exchange offers or business combinations. Use F-7 for pro rata rights distributions; use F-80 for cross-border exchange offers and combinations.
A securities-offering form for non-convertible investment-grade debt or preferred of eligible Canadian issuers. Narrower in instrument type and entirely outside the M&A context. Largely curtailed by later SEC rule changes; historical filings remain on EDGAR. Form F-9 is included here for completeness.
The broadest MJDS Securities Act form, used for primary or secondary cash offerings of any security type, including shelves. Differs from F-80 in two ways: it registers offerings rather than M&A consideration, and it generally requires U.S. GAAP reconciliation that F-80 does not. Form F-10 is far more common than F-80 but not a substitute.
Not a registration statement. F-X is a short submission appointing a U.S. agent for service of process and is filed alongside virtually every MJDS registration, including F-80. Treat it as a companion record to be joined to the substantive filing, not an alternative dataset.
The MJDS counterpart to Schedule TO for third-party tender offers by Canadian bidders for Canadian targets. It satisfies Williams Act tender-offer disclosure; F-80 (or F-8) registers any new securities offered as consideration. In a cross-border exchange offer, both filings exist for the same deal. Cash-only MJDS tender offers appear in Schedule 14D-1F but not F-80.
The MJDS counterpart for issuer self-tenders (a Canadian issuer buying back its own securities). Structurally analogous to 14D-1F but for self-tenders. Overlaps F-80 only when a self-tender involves issuing new securities as consideration; cash repurchases appear in 13E-4F alone.
The fallback when MJDS is unavailable. Form F-4 registers the same transactional fact pattern as F-80 (securities issued in M&A) but for any foreign private issuer and under full U.S.-style disclosure, typically including U.S. GAAP reconciliation and SEC-form prospectus content rather than Canadian provincial documents. Canadian issuers with U.S. ownership above the MJDS threshold, or whose targets fail MJDS criteria, file F-4 instead. F-4 datasets are far larger and contain materially heavier disclosure.
The U.S. domestic counterpart of F-4 and the most distant comparison. Same transaction type as F-80 (M&A consideration) but disjoint filer populations: Form S-4 is for U.S. domestic registrants, F-80 for eligible Canadian foreign private issuers. A given deal will appear on one regime, not the other. Useful only as the domestic baseline against which MJDS accommodations are measured.
The Form F-80 Files Dataset captures a narrow slice: Securities Act registration statements for exchange offers and business combinations by Canadian issuers using MJDS at the lower public float / U.S. ownership tier, with disclosure carried primarily by Canadian provincial documents and no U.S. GAAP reconciliation. F-7, F-9, and F-10 cover different MJDS transaction types. F-8 covers the same transactions at a different eligibility tier. F-X is procedural. 14D-1F and 13E-4F handle the tender-offer side and complement F-80 when consideration includes new securities. F-4 and S-4 cover the same M&A purpose outside MJDS with substantially heavier disclosure. For any cross-border Canada-U.S. exchange offer at the MJDS mid-tier, F-80 is the primary record and 14D-1F, F-X, and where applicable F-8 or F-4 complete the picture.
Users of the Form F-80 Files Dataset are concentrated in cross-border transactional, analytical, and compliance roles, drawing on the MJDS-specific structure of the corpus.
MJDS-experienced securities lawyers use the corpus as a precedent library when drafting exchange offers and plans of arrangement. They mine prior filings for offer mechanics, lock-ups, withdrawal rights, fractional-share treatment, dual-jurisdiction tax disclosure, and the scoping of counsel opinions and consents. F-80/A amendments are studied for staff comment patterns and to gauge how issuers documented MJDS eligibility (the C$75 million float and U.S. ownership thresholds).
Buy-side and sell-side diligence teams use registered offering and proxy circulars to reconstruct comparable cross-border deal structures: consideration mix, exchange ratios, dissent rights, regulatory conditions, break fees, and second-step squeeze-outs under Canadian corporate statutes.
Bankers structuring stock-for-stock deals involving Canadian issuers benchmark exchange-ratio formulas, collar structures, and timelines from initial F-80 to effectiveness, and use registered consideration descriptions to calibrate fairness analyses for analogous mandates.
Arb desks rely on F-80 and F-80/A filings to confirm registration status, extract minimum tender conditions and exact exchange ratios, and treat amendment cadence as a signal of repricing, extensions, or regulatory friction in spread positions.
Because F-80 financials are presented under Canadian GAAP or IFRS without U.S. GAAP reconciliation, reporting specialists use the corpus to compare home-country treatment of revenue, business combinations, share-based payments, and impairment in a U.S. registration context.
In-house securities counsel at Canadian public companies and their U.S. registration agents use the dataset to confirm exhibit lists, signature blocks, and consents against effective peer filings, and to build internal checklists for MJDS submissions.
Governance analysts use the proxy-style disclosure in F-80 packages to study board approval processes, fairness opinions, voting and lock-up agreements, and insider treatment in cross-border combinations.
Securities-regulation researchers use the bounded F-80 corpus to measure MJDS usage trends, the share of Canadian cross-border deals routed through F-80 versus alternative structures, and the effect of regulatory changes on filing volume.
Engineering teams ingest F-80 metadata and documents into cross-border filings pipelines. The small, well-bounded form taxonomy makes the corpus a useful evaluation set for extraction of exchange-offer terms, eligibility statements, and counsel-opinion language.
Credit analysts use registered financial statements and post-transaction capital-structure descriptions to project pro forma leverage and obligor changes, with debt-related consents and exhibits informing covenant views through the transaction.
The use cases below tie directly to the record contents — the EDGAR header in metadata.json, the Canadian offering or take-over bid circular embedded in the primary HTML payload, and the consents, opinions, and agreement exhibits in documentFormatFiles[].
Securities counsel drafting a Canadian-issuer exchange offer pull the primary HTML payloads across all F-80 records to assemble a precedent set of take-over bid circulars and information circulars. They extract sections on minimum tender conditions, withdrawal rights, fractional-share treatment, second-step compulsory acquisition mechanics, and dual-jurisdiction tax disclosure to reuse drafting patterns and benchmark eligibility recitations against the C$75 million float and 40% U.S. ownership thresholds.
Event-driven analysts join initial F-80 records with their F-80/A amendments under the same 333-… fileNo to build a per-deal timeline. The filedAt timestamp, the [Amend] marker in description, and the textual diff across amendment HTML payloads expose repricings, extension of the offer period, and changes to minimum-tender conditions, feeding directly into spread sizing and roll decisions on active deals.
Capital-markets and diligence teams parse the offering document body to populate a structured table of consideration mix, exchange ratios, collar formulas, break fees, lock-up parties, and regulatory conditions for every F-80 deal since 2002. The entities[] array supplies offeror, subject, and co-registrant CIKs to key the records to other deal datasets; the parsed terms support comparable-deal screens for fairness opinions and pitch books.
Issuer compliance teams and registration agents iterate over documentFormatFiles[] to pull EX-5.1 legality opinions, EX-23.1 auditor consents, and the U.S. agent-for-service signature block from the primary document. Aggregating across the corpus produces a roster of which Canadian and U.S. firms appear on which deals, useful for vendor selection, conflict checks, and benchmarking exhibit lists when preparing a new MJDS submission.
Forensic accountants and reporting specialists use the post-2011 cohort of records to study how Canadian issuers present revenue recognition, business-combination accounting, share-based payments, and impairment under IFRS without U.S. GAAP reconciliation. Because the dataset preserves the financial statements as filed, it serves as a controlled corpus for cross-framework comparison against U.S.-domestic S-4 and F-4 registrants on analogous transactions.
Securities-regulation researchers treat the bounded F-80 corpus as a complete population. Counts of F-80 versus F-80/A records by year, joined to F-8, F-10, and F-4 populations, support studies of MJDS filing volume, the share of Canadian cross-border deals routed through F-80 versus F-4, and responsiveness to rule changes — all driven by formType, filedAt, and the Canadian province encoded in stateOfIncorporation.
Data engineering and RAG teams use the corpus as a small, well-bounded benchmark for document-extraction models. The mix of older plain-text submissions and modern HTML payloads, combined with a stable documentFormatFiles[] inventory and predictable exhibit types, makes it a useful test bed for extracting eligibility statements, exchange ratios, and counsel-opinion language without the volume burden of the broader registration-statement universe.
Dataset Index JSON API: https://api.sec-api.io/datasets/form-f80-files.json
This endpoint returns dataset-level metadata (name, description, last updated timestamp, earliest sample date, total records, total size, form types covered, container format, and included file types), the download URL for the full dataset archive, and the list of all individual container files with their per-container size, record count, updated timestamp, and download URL. Use this endpoint to monitor on a daily basis which containers were updated in the most recent refresh run, then decide which containers to re-download. This endpoint does not require an API key.
Example response:
1
{
2
"datasetId": "1f13365b-9ae0-6a3a-bbcf-d39d215cc7bf",
3
"datasetDownloadUrl": "https://api.sec-api.io/datasets/form-f80-files.zip",
4
"name": "Form F-80 Files Dataset",
5
"updatedAt": "2026-04-16T08:37:47.154Z",
6
"earliestSampleDate": "2002-01-01",
7
"totalRecords": 464,
8
"totalSize": 13652850,
9
"formTypes": ["F-80", "F-80/A"],
10
"containerFormat": "ZIP",
11
"fileTypes": ["TXT", "JSON", "HTML"],
12
"containers": [
13
{
14
"downloadUrl": "https://api.sec-api.io/datasets/form-f80-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-f80-files.zip?token=YOUR_API_KEY
Downloads the complete Form F-80 Files Dataset, covering all F-80 and F-80/A filings from January 2002 to present, as a single ZIP archive. This endpoint requires an API key.
Download Single Container: https://api.sec-api.io/datasets/form-f80-files/2026/2026-03.zip?token=YOUR_API_KEY
Downloads one individual monthly container archive instead of the full dataset, which is useful for incremental updates or fetching only a specific time period. This endpoint requires an API key.
The dataset covers Form F-80 (initial registration statement) and Form F-80/A (amendment), both filed under the Securities Act of 1933 by eligible Canadian foreign private issuers under the U.S.–Canada Multijurisdictional Disclosure System. No other form types are included.
One record is a single EDGAR submission of a Form F-80 or F-80/A registration statement, packaged as a per-accession folder named with the 18-digit, dash-stripped accession number. The folder contains exactly one metadata.json describing the EDGAR header plus the original submission documents (HTML or TXT) preserved with their SGML <DOCUMENT> envelopes. An amendment produces an additional, independently retrievable record under its own accession.
Form F-80 is restricted to Canadian foreign private issuers eligible to use MJDS as a "substantial issuer." The registrant must be incorporated in a Canadian jurisdiction, have a class of securities listed on a designated Canadian exchange (typically TSX or Montreal Exchange) for at least 12 months, have an aggregate equity public float of at least C$75 million, and be making an offer in which less than 40% of the subject class is held by U.S. holders.
Form F-80 is event-driven, not periodic. A filing is triggered when an MJDS-eligible Canadian issuer extends a takeover bid (tender or exchange offer) into the United States, or undertakes a business combination — plan of arrangement, amalgamation, or merger — in which it issues its own securities to security holders of another entity. The initial F-80 must be on file no later than the date the Canadian offering or proxy materials are first sent to U.S. holders.
Both register securities issued as consideration in exchange offers and business combinations, but F-80 is available only to MJDS-eligible Canadian issuers and uses the Canadian provincial offering or proxy document without U.S. GAAP reconciliation. F-4 is the fallback used by any foreign private issuer that does not qualify for MJDS, and it requires full U.S.-style prospectus disclosure typically including U.S. GAAP reconciliation. A given transaction is filed on one regime or the other, not both.
Coverage begins January 1, 2002, which is the practical start of electronic coverage because mandatory EDGAR filing for foreign private issuers, including MJDS filers, took effect in November 2002. Earlier F-80 filings exist only in paper form in SEC historical files and are not included.
The dataset is distributed as a ZIP container archive partitioned by year and month (YYYY/YYYY-MM/<accession>/). Each accession folder contains a metadata.json JSON file plus the original EDGAR submission documents as HTML or TXT files preserved with their SGML <DOCUMENT> envelopes. Image binaries (GRAPHIC documents such as JPEG and GIF) are catalogued in metadata.json but intentionally omitted from the archive.