The Form DFAN14A Files Dataset is a complete corpus of definitive additional proxy soliciting materials filed on EDGAR by non-management parties under Section 14(a) of the Securities Exchange Act of 1934 and Rule 14a-6. Each record is a single EDGAR submission — typically an open letter, fight deck, press release, or ISS response issued by an activist investor, dissident slate, shareholder coalition, or contested-deal counterparty — packaged with its primary HTML cover document, all HTML and PDF exhibits as filed, and a generated metadata.json envelope describing the filing, its document inventory, and the soliciting and subject entities. The dataset spans October 1994 to the present and is delivered as monthly ZIP containers with HTML, PDF, TXT, and JSON content. It captures the dissident side of every U.S. proxy contest of the modern EDGAR era, sitting alongside the registrant-side DEFA14A stream and the primary non-management proxy track (PREN14A, PRRN14A, DFRN14A).
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:
Each record in the dataset is a single EDGAR submission of Form DFAN14A — the "definitive additional" variant of the non-management Schedule 14A track. The EDGAR form code decomposes as: leading D for definitive (as opposed to preliminary, P); FAN for "filed by a non-management" person (as opposed to F for the registrant's own filings); and 14A for Schedule 14A. DFAN14A is not the primary proxy statement of the company holding the meeting — that is the registrant's own DEF 14A — but rather a supplementary soliciting material communication issued independently by a party other than the registrant. Typical content includes open letters to shareholders, press releases, investor-presentation decks, vote-recommendation memos, ISS/Glass Lewis response letters, white papers attacking the incumbent board's strategy, deadline-extension or correction notices, and rebuttal statements issued during a contested vote.
The unit of observation is the accession: an 18-digit zero-padded EDGAR accession number under which a soliciting party (typically an activist investor, dissident shareholder group, hedge fund, or coordinating "participants" group) deposits one round of definitive supplemental proxy material directed at the holders of a specific registrant's voting securities. On disk, each record is an accession-number folder — for example 000110465925076166/ — sitting inside a monthly ZIP container named YYYY-MM/. Inside the folder are the SGML-wrapped HTML and binary PDF documents EDGAR received for the submission, plus the metadata.json envelope. Image binaries referenced by the HTML are deliberately excluded; everything else from the submission is preserved as filed. The dataset's earliest sample date is October 1, 1994, tracking the phase-in of mandatory electronic filing of proxy soliciting materials between 1993 and 1996.
A DFAN14A submission almost always presupposes a prior non-management filing (a PREN14A, DEFN14A, or earlier DFAN14A) on the same EDGAR file number and meeting cycle; the "additional" qualifier means the materials supplement an already-launched solicitation rather than initiate it. Each filing rides on the Schedule 14A cover-page framework prescribed by Rule 14a-101: the "UNITED STATES / SECURITIES AND EXCHANGE COMMISSION / Washington, D.C. 20549 / SCHEDULE 14A" header, the "Filed by the Registrant / Filed by a Party other than the Registrant" checkbox block (always the latter for DFAN14A), the materials-classification checkboxes, the registrant identification block, the filer identification block, and a fee-payment statement. After this regulatory cover, the substantive communication appears as free-form narrative, often accompanied by one or more exhibits.
A record contains three concentric layers, mapped to a flat folder on disk:
metadata.json providing accession-level identifiers, the filing timestamp, EDGAR URLs, an exhaustive document inventory, and structured entity records distinguishing soliciting parties from the subject company.sequence: "1", carrying the Schedule 14A cover page and the body of the additional solicitation message.These three layers correspond to: metadata.json; the primary *_dfan14a.htm (filer-chosen filename, no fixed pattern — observed forms include tm2523113d2_dfan14a.htm, dfan14a10318025_08272025.htm, smolyansky_dfan14a.htm, p25-1723dfan14a.htm, dfan14a180degree-08152025.htm); and any further *.htm or *.pdf exhibits.
metadata.json envelopemetadata.json is a flat JSON object emitted exactly once per accession. Its top-level keys describe the filing and its components:
formType — always the string "DFAN14A" for records in this dataset.accessionNo — EDGAR accession number in dashed canonical form, e.g. "0001104659-25-076166". The 18-digit folder name is the same value with dashes removed.description — the static EDGAR description of the form type, e.g. "Form DFAN14A - Additional definitive proxy soliciting materials filed by non-management and [Rule 14(a)(12)](https://www.ecfr.gov/current/title-17/chapter-II/part-240/subpart-A/subject-group-ECFR8c9733e13b955d6/section-240.14a-12) material".filedAt — ISO-8601 timestamp with timezone offset capturing the EDGAR acceptance time, e.g. "2025-08-11T16:16:54-04:00".linkToFilingDetails — URL of the primary DFAN14A document on EDGAR.linkToTxt — URL of the full-submission .txt archive on EDGAR. The complete-submission .txt itself is not redistributed inside the ZIP; this field gives the canonical EDGAR address.linkToHtml — URL of the EDGAR filing index page (*-index.htm).linkToXbrl — empty string for DFAN14A records.documentFormatFiles[] — one entry per item EDGAR catalogued for the submission (see below).entities[] — one entry per filer or subject company (see below).seriesAndClassesContractsInformation[] — empty array for this form type.dataFiles[] — empty array for this form type.id — a 32-character hexadecimal internal record identifier.documentFormatFiles[]Each element describes one EDGAR-indexed attachment with these fields:
sequence — string EDGAR sequence number; the trailing complete-submission entry uses a single space " ".documentUrl — absolute www.sec.gov/Archives/... URL.type — EDGAR document-type code such as DFAN14A, EX-99.1, GRAPHIC, or " " for the complete-submission file.description — free-text label such as "DFAN14A", "EXHIBIT 99.1", "GRAPHIC", "DEFINITIVE ADDITIONAL MATERIALS", or "Complete submission text file". Occasionally omitted on minimal entries.size — string-encoded byte count.The basename of documentUrl matches the on-disk filename for every non-image, non-trailing entry. Image entries (type: "GRAPHIC") are listed for completeness but their bytes are not shipped, and the trailing complete-submission entry is likewise advertised in metadata but not bundled.
entities[]Two or more entity objects per record, distinguished by a parenthetical suffix on companyName:
(Filed by) entities — the soliciting parties. The filer is frequently an individual activist or a small fund, and these objects typically carry only cik, companyName, and type. When several cooperating participants file jointly (for example an individual plus an L.P. plus a general partner), multiple (Filed by) entries appear; each is an independent CIK within the meaning of a Rule 14a-1(l) "group of participants".(Subject) entity — the registrant whose shareholders are being solicited. This object carries the full EDGAR header set: cik, companyName, fiscalYearEnd, stateOfIncorporation, act (typically "34"), fileNo, irsNo, sic (with both code and label, e.g. "6500 Real Estate", "2020 Dairy Products", "6035 Savings Institution, Federally Chartered"), filmNo, and type. The subject entity is the reliable join key to other EDGAR datasets keyed on CIK or SIC.Order between (Filed by) and (Subject) entries is not guaranteed — consumers must split the array by suffix rather than by position.
The primary document (type: "DFAN14A", sequence: "1") is an HTML file wrapped in EDGAR's SGML document envelope. The wrapper precedes the <HTML> body and uses the canonical EDGAR header tags:
1
<DOCUMENT>
2
<TYPE>DFAN14A
3
<SEQUENCE>1
4
<FILENAME>...
5
<DESCRIPTION>DFAN14A
6
<TEXT>
7
<HTML>...</HTML>
8
</TEXT>
9
</DOCUMENT>
Inside the HTML body, the document follows the Schedule 14A cover-page convention — "UNITED STATES / SECURITIES AND EXCHANGE COMMISSION / Washington, D.C. 20549 / SCHEDULE 14A / Proxy Statement Pursuant to Section 14(a) of the Exchange Act of 1934 / Rule 14a-101" — followed by:
Below this cover, the substantive solicitation text occupies the remainder of the document — most often a press release, an open letter, a vote-recommendation memo, a slide narrative, or a procedural extension notice. There is no SEC-mandated section structure for the body itself: lengths range from a single paragraph to a multi-hundred-page presentation, and headings are filer-chosen.
Exhibits arrive as additional documents inside the same accession folder, each carrying its own SGML wrapper when in HTML form:
EX-99.1, EX-99.2, etc. (press releases, investor decks, presentation transcripts, supporting analyses). They sit beside the primary HTML with filer-chosen filenames such as tm2523113d2_ex99-1.htm, wrapped in an envelope of the form <DOCUMENT> <TYPE>EX-99.1 <SEQUENCE>2 <FILENAME>... <DESCRIPTION>EXHIBIT 99.1 <TEXT> <HTML>...</HTML> </TEXT> </DOCUMENT>. Investor-deck exhibits frequently render slides as embedded JPG references; because the dataset omits image binaries, only the text of speaker-note overlays, slide captions, and surrounding boilerplate survives as parseable content in those exhibits.ex1todfan14a10318025_082725.pdf) sit next to the HTML cover with no SGML wrapper at the file level. They begin with the standard %PDF- magic bytes and may run from a one-page letter to multi-megabyte presentation decks. Filer indexing of PDFs is inconsistent: some declare them in documentFormatFiles with the proper exhibit type (EX-99.1, etc.), while others reuse the form-type label DFAN14A for the exhibit. Distinguishing the cover document from a PDF exhibit therefore requires combining sequence (the cover is always sequence: "1") with file extension and filename rather than relying on type alone.Each record contains, as filed: the primary DFAN14A HTML cover and substantive communication; every HTML exhibit declared on the EDGAR submission with its full SGML wrapper intact; every PDF exhibit as a true binary; and the metadata.json envelope summarizing the filing and indexing every document. The set is sufficient to reconstruct the textual content of the soliciting materials, the identity of every filing participant, the identity and classification of the subject registrant, the EDGAR acceptance timestamp, the EDGAR document inventory, and the canonical URLs for cross-referencing the original filing on www.sec.gov.
Three categories of content advertised in documentFormatFiles are not redistributed inside the per-accession folder:
type: "GRAPHIC") — JPG, GIF, and PNG files referenced by <IMG SRC="..."> tags in the HTML are listed in metadata but their bytes are excluded by design. References inside the HTML resolve to filenames absent from the ZIP; the corresponding documentUrl in metadata is the only path back to the binary on EDGAR..txt archive — the trailing documentFormatFiles entry (sequence " ", type " ", description "Complete submission text file") advertises EDGAR's full-submission .txt (e.g. 0001104659-25-076166.txt) but that aggregate is not bundled; linkToTxt provides the URL.Form DFAN14A as a regulatory construct has a relatively stable scope: it is the definitive additional-materials channel for non-management Schedule 14A solicitation, governed since 1992 by the Rule 14a-6(b) framework that distinguishes preliminary from definitive proxy materials and additional from initial filings. Across the dataset's range — October 1994 to present — the core required cover-page content (Schedule 14A header, Filed-by checkbox, registrant identification, filer identification, materials-classification checkboxes, fee-payment statement) has remained largely intact.
Several rule changes are nevertheless visible in the body content of records across decades:
The "additional" character of DFAN14A means that, beyond the Schedule 14A cover, the body remains highly heterogeneous: there is no SEC-mandated section structure for the substantive communication, and content varies from a single-paragraph extension notice to a multi-hundred-page presentation deck.
The format trajectory of DFAN14A records over the 1994–present span tracks EDGAR's general modernization:
<DOCUMENT> / <TYPE> / <SEQUENCE> / <FILENAME> / <DESCRIPTION> / <TEXT> header format is already in use, but the body inside <TEXT>...</TEXT> is monospaced text rather than HTML, with hand-formatted column alignment and capital-letter section headings.<P>, <B>, table tags) and later with full styled HTML.style attributes, and complex tables. Exhibits begin to appear regularly as separate HTML attachments (EX-99.*) or as binary PDFs.<BODY STYLE="font: 10pt Times New Roman, Times, Serif">) plus one or more HTML or PDF exhibits, with images carried separately as GRAPHIC attachments. PDFs are increasingly used for slide decks and lengthy investor presentations; HTML is preferred for press releases and shareholder letters.The SGML envelope (<DOCUMENT>...</DOCUMENT> with header tags before <TEXT>) persists across the entire timeline as the at-rest packaging convention for every HTML/text document, even after the body content shifted from ASCII to HTML.
YYYY-MM/<accessionNo>/metadata.json. The 18-digit folder name maps to the accessionNo value with dashes inserted at positions 10 and 13.documentFormatFiles is the canonical list of what EDGAR catalogued. To get the files actually present on disk, filter to entries where type is not "GRAPHIC" and sequence is not the trailing space-only entry, then take the basename of documentUrl as the on-disk filename.entities[] by the (Filed by) / (Subject) suffix on companyName, not by array index. The subject entry carries the rich identification block (sic, stateOfIncorporation, fileNo, irsNo, filmNo) for downstream joins; filer entries are usually identification-only.<DOCUMENT>, <TYPE>, <SEQUENCE>, <FILENAME>, <DESCRIPTION>, <TEXT>) and ends with </TEXT></DOCUMENT>. Strip these before passing the body to an HTML parser if a clean DOM is required.<IMG SRC="..."> reference in a DFAN14A or exhibit HTML resolves to a filename absent from the dataset. Either suppress images during rendering or rewrite src to the EDGAR documentUrl listed in metadata. Investor-deck exhibits in particular may render as nearly-blank HTML once images are removed; the textual signal then comes from speaker notes and embedded captions, not from the slide imagery itself.type: "DFAN14A" rather than an EX-* code. Use sequence: "1" plus filename to identify the cover document; treat anything with sequence > 1 as an exhibit regardless of how the type field is set.(Filed by) entries on a single record indicate a coordinated solicitation by a Rule 14a-1(l) "group of participants". Each such entry is a distinct filer with its own CIK; downstream pipelines that key on a single filer must decide whether to expand or aggregate these.seriesAndClassesContractsInformation[] and dataFiles[] are reliably empty for DFAN14A and can be treated as always-empty by consumer code./A suffix in EDGAR's contested-proxy track; corrections, clarifications, or follow-ups arrive as separate DFAN14A filings under new accession numbers, each appearing as an independent record in the dataset. Linking a corrective filing to its predecessor requires matching on subject CIK plus EDGAR fileNo plus the meeting cycle implied by filedAt.Each record is a definitive additional proxy soliciting communication submitted to EDGAR by a non-management soliciting party. The filer is never the registrant. The "N" in DFAN14A signals "non-management": the filing party is an outside participant in a proxy solicitation under Regulation 14A, conducting its own campaign in parallel with — and almost always in opposition to — the issuer's solicitation.
Typical DFAN14A filers include:
Proxy solicitors, counsel, and filing agents may prepare and transmit the materials, but the legal filer of record is the soliciting participant. When multiple persons act jointly (a fund, its general partner, its investment manager, principals, and nominees), a single DFAN14A is typically filed on behalf of the full group, with all participants identified per Schedule 14A.
DFAN14A exists under Section 14(a) of the Securities Exchange Act of 1934, which prohibits proxy solicitation in contravention of SEC rules, and under Regulation 14A (Rules 14a-1 through 14a-21). The relevant pieces:
Within EDGAR's proxy form codes, DFAN14A sits alongside PREN14A (preliminary non-management proxy statement), DEFN14A (definitive non-management proxy statement), and DEFA14A (the management-side counterpart used by the registrant for its own additional soliciting materials).
A DFAN14A is event-driven, not periodic. Each separate piece of additional soliciting material disseminated by the non-management party after its initial proxy contest materials generally generates its own DFAN14A. Typical triggering communications include:
The substantive trigger is constant: an outside party is communicating with the registrant's shareholders in a manner reasonably calculated to influence their vote, after its solicitation has reached the definitive stage. There is no calendar cadence and no fixed count per campaign — a heavily contested fight may produce dozens of DFAN14A filings over several weeks; a "vote no" campaign may produce one or two.
The underlying votes most commonly driving DFAN14A activity are contested director elections, opposed M&A transactions, charter and bylaw amendments, targeted say-on-pay votes, Rule 14a-8 proposals where the proponent solicits beyond the issuer's proxy, and reincorporation or recapitalization votes.
Rule 14a-6(b) sets a single deadline: filing must occur no later than the date the materials are first sent or given to security holders. Unlike a preliminary proxy statement, additional soliciting materials are not subject to a staff review window — they may be disseminated and filed contemporaneously, and in practice are typically filed the same day the material is first published, mailed, emailed, advertised, or otherwise made available. New communications are generally filed as fresh DFAN14A submissions rather than as amendments to prior ones.
Section 14(a) and the proxy rules date to the Exchange Act of 1934 and the Securities and Exchange Commission's 1935–1938 rulemakings. Mandatory electronic filing of proxy soliciting materials was phased in on EDGAR between 1993 and 1996; the earliest DFAN14A records in this dataset begin in October 1994. Pre-EDGAR contested solicitation materials exist in the SEC's paper archives but are not part of this dataset.
DFAN14A sits inside the Section 14(a) proxy solicitation family. The forms below cover overlapping subject matter but each diverges from DFAN14A on a single, specific dimension: filer side, lifecycle stage, primary vs. supplemental role, regulatory regime, or content type.
Same lifecycle (definitive) and same role (additional soliciting material) as DFAN14A. The only difference is filer side: DEFA14A comes from the issuer supplementing its DEF 14A; DFAN14A comes from the dissident or non-management solicitor. In a contested vote the two streams run in parallel and must be merged for a complete picture of the contest.
Same filer side as DFAN14A (non-management). Differs on lifecycle stage and role: PREN14A is the preliminary primary solicitation document, subject to SEC review before becoming definitive. DFAN14A is definitive supplemental material released after clearance.
Non-management filer, like DFAN14A. Differs on lifecycle stage: PRRN14A is preliminary (revising an earlier preliminary statement and potentially subject to further comment), while DFAN14A is definitive. It is also a revision of the primary statement, not supplemental material.
The closest neighbor and most easily confused. Same filer side and same lifecycle as DFAN14A. Differs on document role: DFRN14A is the formal proxy statement itself (in revised form), carrying the full required disclosure set — participants, ownership, proposals, voting procedures, proxy card. DFAN14A is supplemental content layered on top: rebuttal letters, ISS/Glass Lewis announcements, fight letters, investor decks. To reconstruct the dissident's formal position, the primary non-management proxy (PREN14A/PRRN14A/DFRN14A line) is required; DFAN14A alone gives the running commentary, not the foundation document.
Differs from DFAN14A on both filer side and document role: DEF 14A is the registrant's principal, comprehensive proxy statement, with structured disclosure (compensation tables, beneficial ownership, audit committee, formal vote items). DFAN14A is fragmentary, opposition-prepared supplemental material focused on contested points. They share only the underlying shareholder meeting.
Differs on regulatory regime. Schedule 14N covers shareholder director nominations made through the registrant's own proxy card under proxy access rules. DFAN14A is a Schedule 14A filing tied to a competing proxy card and full solicitation campaign. The two regimes address shareholder-driven board change through structurally different mechanisms; their filer populations are largely disjoint.
Differs on content type and statutory basis. 13D is a Section 13(d) ownership disclosure (>5% non-passive stake, source of funds, purpose, plans). DFAN14A is a Section 14(a) solicitation document directed at fellow shareholders. They frequently co-occur in activist campaigns but answer different questions: 13D shows who is accumulating and why; DFAN14A shows what they are telling other holders to win the vote.
DFAN14A occupies a precise intersection: definitive, supplemental, non-management, Section 14(a) solicitation, traditional proxy contest. Each related form flips exactly one of those axes — DEFA14A flips filer side, PREN14A and PRRN14A flip lifecycle stage, DFRN14A flips document role, DEF 14A flips filer side and role, Schedule 14N flips regulatory regime, Schedule 13D flips content type. For the dissident's running solicitation narrative after a definitive proxy is in place, DFAN14A is the central source; for management's voice, the dissident's formal disclosures, or ownership stakes, the adjacent datasets are required complements rather than substitutes.
The roles below each draw on specific parts of the record: the soliciting text body, the (Filed by) participant array, filing timestamps, and exhibits disclosing arrangements and understandings.
Analysts at activist funds pull prior DFAN14A filings against comparable issuers to benchmark fight-letter argumentation, board-fitness frameworks, and capital-allocation critiques. They read the soliciting text body for tested lines of attack, parse the (Filed by) array to map co-investors, nominees, and advisers aligned in past campaigns, and use filing timestamps to model escalation cadence (days from first letter to attack deck, response lag to a registrant DEFA14A). Output: a campaign plan or pitch deck grounded in observed precedent.
M&A and disclosure counsel on both sides use the filings to draft compliant materials and to attack opponents'. Dissident counsel mirrors prior participant tables, ownership disclosures, and Item 4 statements of arrangements; incumbent counsel reads new dissident filings on submission to spot Rule 14a-9 misstatements, omitted participants, or undisclosed group arrangements. Exhibits documenting consulting agreements, indemnification, and nominee compensation are the primary evidence base for complaints to staff.
Solicitors and IR firms mine the soliciting text and timestamp sequence to build campaign calendars, sample-letter libraries, and Q&A templates. The (Filed by) array identifies which solicitors and advisers have historically appeared on which side, supporting competitive pitches to issuers facing an activist threat.
Stewardship analysts at large asset managers and pension funds read every dissident letter and presentation alongside the registrant's defense to form vote rationales on contested ballots. They focus on board-composition arguments, capital-allocation analyses, and the participant disclosures that reveal the activist's holding period and economic exposure. The record supports per-meeting vote memos and aggregate contested-vote reporting.
Proxy advisers consume the soliciting body as core source material for contested-election recommendation reports. They cross-reference the (Filed by) array against incumbent disclosures and build longitudinal campaign databases for sector studies and policy research.
Event-driven analysts monitor new DFAN14A filings within minutes of submission, classify the soliciting party's position (oppose deal, vote-no, replace board) from the text body, and re-size positions on target, acquirer, or related instruments. Timestamps inform deal-break and dissident-win probabilities; the participant array flags new co-investors joining a campaign.
Defense bankers, counsel, and crisis-communications advisers study prior campaigns by the same activist (identified via the (Filed by) array) to anticipate attack vectors and draft counter-narratives. Exhibits and cited operating data reveal which board minutes or internal materials the dissident is likely to demand or quote next.
Researchers use the structured (Filed by) array and the 1994-onward span to link activists, targets, nominees, and outcomes for event studies on abnormal returns, settlement dynamics, and downstream governance change. The text body supports NLP work on fight-letter rhetoric and the diffusion of ESG-themed arguments.
Reporters use the soliciting body for direct quotes, the (Filed by) array to identify aligned parties not yet named in press releases, timestamps to reconstruct campaign chronology, and exhibits for evidence of consulting arrangements or undisclosed groupings. Output: breaking coverage of campaign escalations and investigative pieces on activist networks.
Quant teams ingest the full corpus to extract features for campaign intensity, dissident sentiment, escalation cadence, and participant overlap. NLP over the soliciting text classifies campaign type (board change, M&A opposition, strategic review, ESG) and extracted demands; the (Filed by) array, timestamps, and exhibit metadata feed factor models and event-study pipelines.
Compliance teams at brokers, custodians, and regulators screen the participant identities, ownership disclosures, and arrangement exhibits for intersections with client lists, potential undisclosed group activity, and solicitations around securities under heightened review.
The common need across these roles is a complete, time-stamped, machine-readable record of the dissident side of every U.S. proxy solicitation since 1994: the soliciting text for argument, the (Filed by) array for participant identity, timestamps for sequencing, and exhibits for arrangements.
Specific workflows the Form DFAN14A Files Dataset supports end-to-end.
Reconstructing campaign timelines for a contested vote. Filter records by (Subject) CIK and EDGAR fileNo, sort by filedAt, and merge with the matching DEFA14A stream to produce a unified dissident-vs-registrant chronology of letters, decks, and ISS responses. Output: a per-meeting timeline used in vote memos, deal-break models, and post-mortems on solicitation cadence.
Building an activist messaging library. Strip the SGML wrapper from each sequence: "1" HTML body and the EX-99.* exhibits, tag by SIC and campaign theme (board change, M&A opposition, strategic review, ESG), and store the cleaned text as a searchable corpus. Activist-side strategists query it for prior fight-letter language against comparable issuers; defense advisers query it for known attack patterns from a specific filer.
Mapping participant networks across campaigns. Partition entities[] by the (Filed by) suffix, collect every co-filing CIK pair across the dataset, and join against (Subject) CIKs to build an edge list of activists, nominees, advisers, and L.P./G.P. structures appearing together in Rule 14a-1(l) participant groups. Output: a network graph powering pitch decks, journalism on activist syndicates, and academic event studies.
Real-time event-driven signals on new filings. Poll the dataset for new accessions, parse the body for vote-recommendation verbs ("vote AGAINST", "WITHHOLD", "tender no shares") and Rule 14a-19 universal proxy card mechanics, and emit a structured signal keyed on (Subject) CIK with filedAt as the event time. Feeds merger-arb book sizing, deal-break probability features, and dissident-win priors.
Quantifying escalation cadence as an NLP feature. Compute time deltas between successive DFAN14A filedAt timestamps on the same fileNo, combine with body-length growth, sentiment scoring, and exhibit-count trajectory across the campaign, and use the resulting series as a campaign-intensity factor in event studies and abnormal-return models.
Surveillance for undisclosed groups and arrangements. Cross-reference (Filed by) CIKs against the participant tables and Item 4 arrangement disclosures inside the body and exhibits, then flag DFAN14A filings where new economic actors appear without a corresponding Schedule 13D/G amendment or where consulting and indemnification exhibits suggest undisclosed coordination. Used by incumbent counsel preparing 14a-9 complaints to staff.
Drafting vote rationales at stewardship desks. For each contested ballot in scope, pull the full DFAN14A series for the meeting plus the registrant's DEFA14A counterstream, extract board-composition arguments and capital-allocation analysis from the bodies and PDF decks, and pair with (Filed by) ownership and nominee disclosures. Output: a per-meeting vote memo with citations to specific accession numbers and EDGAR linkToFilingDetails URLs.
Longitudinal research on contested-proxy rhetoric. Use the 1994-onward span and the rule-change inflection points (1999 plain English, 2009 e-proxy, 2022 universal proxy) as natural-experiment boundaries, run NLP over the soliciting text bodies, and quantify shifts in fight-letter language, ESG framing, and nominee qualification disclosures. Supports law-review and finance-journal papers on activism dynamics.
Dataset Index JSON API: https://api.sec-api.io/datasets/form-dfan14a-files.json
This endpoint returns the dataset's metadata, including its name, description, last update timestamp, earliest sample date (1994-10-01), covered form types (DFAN14A), container format (ZIP), and content file types (TXT, JSON, PDF, HTML). It also returns the full dataset download URL and the complete list of individual container files, with per-container size, record count, updated timestamp, and download URL. Poll this endpoint to detect which monthly containers changed in the most recent refresh and fetch only those incrementally. No API key is required to access this endpoint.
1
{
2
"datasetId": "1f13365b-9ae0-690e-9d5c-5035021e3072",
3
"datasetDownloadUrl": "https://api.sec-api.io/datasets/form-dfan14a-files.zip",
4
"name": "Form DFAN14A Files Dataset",
5
"updatedAt": "2026-05-07T02:50:16.984Z",
6
"earliestSampleDate": "1994-10-01",
7
"totalRecords": 18760,
8
"totalSize": 2364073934,
9
"formTypes": ["DFAN14A"],
10
"containerFormat": "ZIP",
11
"fileTypes": ["TXT", "JSON", "PDF", "HTML"],
12
"containers": [
13
{
14
"downloadUrl": "https://api.sec-api.io/datasets/form-dfan14a-files/2026/2026-05.zip",
15
"key": "2026/2026-05.zip",
16
"size": 13818783,
17
"records": 154,
18
"updatedAt": "2026-05-07T02:50:16.984Z"
19
}
20
]
21
}
Download Entire Dataset: https://api.sec-api.io/datasets/form-dfan14a-files.zip?token=YOUR_API_KEY
Downloads the full Form DFAN14A archive as a single ZIP containing all monthly containers from October 1994 to present. This endpoint requires an API key.
Download Single Container: https://api.sec-api.io/datasets/form-dfan14a-files/2026/2026-05.zip?token=YOUR_API_KEY
Downloads one monthly container ZIP, which is useful for incremental updates or backfilling a specific month without retrieving the entire dataset. Replace the year and month path segments to target a different period. This endpoint requires an API key.
One record is a single EDGAR submission of Form DFAN14A — one round of definitive additional proxy soliciting materials filed under an 18-digit accession number by a non-management party. Each record is delivered as an accession-number folder containing the primary DFAN14A HTML cover, any HTML or PDF exhibits, and a metadata.json envelope, all sitting inside a monthly ZIP container named YYYY-MM/.
The filer is always a non-management soliciting party — most commonly an activist hedge fund, dissident director slate, shareholder coalition, contested-deal counterparty, founder or large individual holder, or institutional investor running a "vote no" or governance campaign. The registrant never files DFAN14A; its equivalent supplemental materials are filed as DEFA14A.
Rule 14a-6(b) requires definitive additional soliciting materials to be filed with the Commission no later than the date the materials are first sent or given to security holders. Additional materials are not subject to a staff review window, so they are typically filed the same day they are first published, mailed, emailed, advertised, or otherwise made available.
The dataset begins on October 1, 1994 — the earliest DFAN14A record on EDGAR after mandatory electronic filing of proxy soliciting materials was phased in between 1993 and 1996 — and extends to the present, with monthly container ZIPs added on an ongoing refresh cadence.
A record contains a primary HTML cover document at sequence: "1" wrapped in EDGAR's SGML envelope, zero or more HTML exhibits (typically EX-99.1, EX-99.2, etc.) with their own SGML wrappers, and zero or more binary PDF exhibits. The metadata.json envelope inventories every document. Image binaries (type: "GRAPHIC") and the aggregated complete-submission .txt archive are advertised in metadata but excluded from the per-accession folder.
DEFA14A is the registrant-side equivalent of DFAN14A: same definitive supplemental role, but filed by the issuer rather than by the dissident. DFRN14A, despite the similar code, is the dissident's formal primary proxy statement in revised form, carrying the full Schedule 14A disclosure set; DFAN14A is supplemental commentary layered on top of that primary statement (open letters, fight decks, ISS responses).
The dataset is delivered as monthly ZIP containers and refreshed on an ongoing cadence. The Dataset Index JSON API at https://api.sec-api.io/datasets/form-dfan14a-files.json returns per-container updatedAt timestamps, record counts, and sizes — poll it to detect which monthly containers changed in the most recent refresh and download only those incrementally.
.txt file included?No. Image attachments referenced by <IMG SRC="..."> tags in the HTML are listed in documentFormatFiles but their bytes are deliberately excluded from the dataset, and the trailing complete-submission .txt aggregate is similarly advertised in metadata but not bundled. Both can be retrieved from EDGAR directly using the documentUrl and linkToTxt fields in metadata.json.