Form DRSLTR Files Dataset

The Form DRSLTR Files Dataset packages every EDGAR submission of form type DRSLTR — issuer correspondence transmitted to the staff of the SEC's Division of Corporation Finance during confidential, non-public review of a draft registration statement. Each record captures one accession number: the structured EDGAR header plus the textual letter(s) the issuer (almost always through outside securities counsel) submitted under that accession. The form was introduced into EDGAR on October 1, 2012 alongside Form DRS to operationalize the JOBS Act's confidential-submission regime, so the dataset's earliest sample date is 2012-10-01 and coverage runs forward from there. Records are delivered as monthly ZIP containers holding HTML, JSON, TXT, and PDF artifacts, with a uniform metadata.json envelope describing the EDGAR header for each filing.

Update Frequency
Daily
Updated at
2026-05-09
Earliest Sample Date
2012-10-01
Total Size
158.9 MB
Total Records
6,546
Container Format
ZIP
Content Types
HTML, JSON, TXT, PDF
Form Types
DRSLTR

Dataset APIs

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:

Dataset Files

156 files · 158.9 MB
Download All
2025-09.zip11.0 KB2 records
2025-08.zip160.9 KB47 records
2025-07.zip136.9 KB49 records
2025-06.zip240.5 KB38 records
2025-05.zip272.2 KB35 records
2025-04.zip240.3 KB48 records
2025-03.zip358.2 KB99 records
2025-02.zip264.4 KB71 records
2025-01.zip345.9 KB83 records
2024-12.zip394.8 KB95 records
2024-11.zip338.0 KB87 records
2024-10.zip323.4 KB44 records
2024-09.zip533.4 KB69 records
2024-08.zip495.0 KB121 records
2024-07.zip6.0 MB53 records
2024-06.zip426.2 KB50 records
2024-05.zip421.2 KB99 records
2024-04.zip3.0 MB43 records
2024-03.zip337.3 KB43 records
2024-02.zip319.2 KB44 records
2024-01.zip314.6 KB41 records
2023-12.zip429.2 KB115 records
2023-11.zip344.2 KB45 records
2023-10.zip253.4 KB36 records
2023-09.zip430.2 KB109 records
2023-08.zip296.0 KB85 records
2023-07.zip353.8 KB81 records
2023-06.zip427.4 KB53 records
2023-05.zip386.5 KB50 records
2023-04.zip293.3 KB38 records
2023-03.zip327.5 KB46 records
2023-02.zip409.9 KB50 records
2023-01.zip372.0 KB46 records
2022-12.zip425.7 KB58 records
2022-11.zip327.2 KB45 records
2022-10.zip369.1 KB41 records
2022-09.zip412.7 KB46 records
2022-08.zip407.8 KB50 records
2022-07.zip380.8 KB42 records
2022-06.zip236.4 KB30 records
2022-05.zip275.6 KB39 records
2022-04.zip313.8 KB39 records
2022-03.zip308.1 KB41 records
2022-02.zip331.3 KB46 records
2022-01.zip323.9 KB41 records
2021-12.zip392.1 KB92 records
2021-11.zip244.3 KB31 records
2021-10.zip334.4 KB47 records
2021-09.zip56.6 MB190 records
2021-08.zip659.1 KB83 records
2021-07.zip529.2 KB69 records
2021-06.zip27.8 MB91 records
2021-05.zip754.6 KB88 records
2021-04.zip522.8 KB62 records
2021-03.zip481.9 KB67 records
2021-02.zip516.4 KB70 records
2021-01.zip456.5 KB64 records
2020-12.zip540.4 KB70 records
2020-11.zip586.5 KB58 records
2020-10.zip380.0 KB48 records
2020-09.zip435.2 KB62 records
2020-08.zip396.4 KB52 records
2020-07.zip248.1 KB39 records
2020-06.zip446.9 KB65 records
2020-05.zip323.4 KB47 records
2020-04.zip209.4 KB33 records
2020-03.zip251.6 KB35 records
2020-02.zip249.4 KB41 records
2020-01.zip338.6 KB46 records
2019-12.zip216.9 KB31 records
2019-11.zip225.9 KB33 records
2019-10.zip229.9 KB32 records
2019-09.zip287.2 KB43 records
2019-08.zip441.5 KB57 records
2019-07.zip261.9 KB36 records
2019-06.zip3.0 MB48 records
2019-05.zip374.7 KB48 records
2019-04.zip393.1 KB48 records
2019-03.zip361.8 KB41 records
2019-02.zip227.4 KB35 records
2019-01.zip97.4 KB14 records
2018-12.zip357.1 KB34 records
2018-11.zip342.5 KB32 records
2018-10.zip262.9 KB38 records
2018-09.zip290.2 KB33 records
2018-08.zip412.3 KB60 records
2018-07.zip450.0 KB39 records
2018-06.zip3.5 MB50 records
2018-05.zip450.9 KB59 records
2018-04.zip13.0 MB32 records
2018-03.zip412.9 KB51 records
2018-02.zip295.6 KB40 records
2018-01.zip316.1 KB39 records
2017-12.zip253.3 KB38 records
2017-11.zip213.3 KB28 records
2017-10.zip301.0 KB35 records
2017-09.zip528.9 KB53 records
2017-08.zip464.0 KB46 records
2017-07.zip362.8 KB36 records
2017-06.zip266.2 KB31 records
2017-05.zip314.7 KB30 records
2017-04.zip262.1 KB24 records
2017-03.zip342.3 KB32 records
2017-02.zip262.7 KB31 records
2017-01.zip435.4 KB29 records
2016-12.zip186.3 KB26 records
2016-11.zip153.6 KB19 records
2016-10.zip120.4 KB16 records
2016-09.zip188.9 KB24 records
2016-08.zip194.2 KB20 records
2016-07.zip163.7 KB18 records
2016-06.zip215.9 KB24 records
2016-05.zip185.8 KB22 records
2016-04.zip184.7 KB22 records
2016-03.zip224.9 KB24 records
2016-02.zip95.4 KB12 records
2016-01.zip92.6 KB10 records
2015-12.zip238.8 KB30 records
2015-11.zip113.8 KB15 records
2015-10.zip224.7 KB26 records
2015-09.zip172.4 KB19 records
2015-08.zip257.1 KB25 records
2015-07.zip132.9 KB14 records
2015-06.zip399.2 KB21 records
2015-05.zip164.8 KB23 records
2015-04.zip286.5 KB27 records
2015-03.zip216.9 KB22 records
2015-02.zip93.9 KB15 records
2015-01.zip168.4 KB16 records
2014-12.zip270.4 KB32 records
2014-11.zip357.8 KB31 records
2014-10.zip313.4 KB24 records
2014-09.zip336.9 KB30 records
2014-08.zip518.9 KB44 records
2014-07.zip288.1 KB24 records
2014-06.zip529.8 KB33 records
2014-05.zip306.9 KB22 records
2014-04.zip220.6 KB22 records
2014-03.zip284.3 KB22 records
2014-02.zip306.7 KB24 records
2014-01.zip341.3 KB31 records
2013-12.zip160.4 KB21 records
2013-11.zip130.0 KB11 records
2013-10.zip76.6 KB12 records
2013-09.zip551.8 KB25 records
2013-08.zip248.4 KB20 records
2013-07.zip241.0 KB27 records
2013-06.zip245.7 KB19 records
2013-05.zip520.6 KB26 records
2013-04.zip210.5 KB19 records
2013-03.zip230.1 KB22 records
2013-02.zip190.7 KB14 records
2013-01.zip172.2 KB12 records
2012-12.zip186.3 KB14 records
2012-11.zip69.1 KB5 records
2012-10.zip2.2 KB1 records

What This Dataset Contains

The dataset covers the full population of publicly visible Form DRSLTR submissions on EDGAR. Form DRSLTR is the EDGAR submission type used to transmit issuer correspondence in connection with a draft registration statement (filed under EDGAR types DRS or DRS/A). Draft registration statements are reviewed confidentially by the Division of Corporation Finance under procedures originally created by Title I of the JOBS Act of 2012 for emerging growth companies, then expanded in July 2017 by Division-of-Corporation-Finance policy to any issuer's first registration statement and to certain follow-on offerings within twelve months of an IPO, with the FAST Act technical amendments of 2018 codifying parts of the expansion. DRSLTR is the channel through which filers submit cover letters announcing a confidential or revised draft, point-by-point responses to staff comment letters, supplemental analyses on accounting or disclosure issues, requests for confidential treatment, and acknowledgments of the public-filing timing requirements.

DRSLTR submissions are themselves nonpublic at acceptance and become available on EDGAR only after the issuer publicly files the registration statement, at which point the entire confidential history (the underlying drafts and the DRSLTR correspondence) is released. The dataset reflects this: it contains only DRSLTR records that have become publicly visible. The underlying letter is almost always a counsel-prepared document on outside-law-firm letterhead, structured as a formal business letter rather than as a regulatory form — there are no numbered Items, no Parts, no prescribed schedules. Its anatomy is dictated by professional convention rather than Regulation S-K.

The dataset is distributed as monthly ZIP containers. Inside each container, every record is a folder named with the eighteen-digit unpunctuated accession number (for example 000114036125033464), bundling a structured metadata.json file together with the correspondence document(s) EDGAR accepted under that accession. File types found across the corpus are HTML, JSON, TXT, and PDF.

Content Structure of a Single Record

What one record represents

One record corresponds to a single EDGAR submission of Form DRSLTR — one accession number, one correspondence filing transmitted by or on behalf of an issuer to the staff of the SEC's Division of Corporation Finance during the confidential, nonpublic review of a draft registration statement. Each record is materialized as a folder named with the eighteen-digit unpunctuated accession number (for example 000114036125033464), placed inside a monthly ZIP container, and bundling a structured metadata file together with the correspondence document(s) that EDGAR accepted under that accession. The unit of analysis is the filing as a whole: every letter, attachment row, filer entity, and submission-level attribute that EDGAR associates with the accession is contained in the one folder.

Folder layout

Each accession folder is organized into two layers. The first is a structured metadata layer, embodied in a single metadata.json file at the root of the folder, which captures the EDGAR header, identifiers, filer entities, and an inventory of every document originally submitted. The second is the document layer: one or more SGML-wrapped HTML files holding the letter text and any textual attachments, with the primary letter conventionally named filename1.htm (the default name EDGAR assigns for DRSLTR letters). Embedded images referenced from the HTML are excluded from the dataset by design, so the document layer contains only machine-readable text artifacts. The file-types found in the dataset are HTML, JSON, TXT, and PDF — covering the full historical range — although in practice the overwhelming majority of records contain only the metadata.json and one HTML letter; older or unusual submissions can carry a TXT-only letter or a PDF attachment instead.

metadata.json

The metadata envelope is uniform across every record; only the contents of array fields vary. The top-level fields are:

  • formType — always the literal string DRSLTR.
  • accessionNo — the punctuated EDGAR accession number, e.g. 0001140361-25-033464.
  • filedAt — ISO 8601 timestamp with Eastern Time offset capturing the EDGAR acceptance time, e.g. 2025-08-29T17:11:54-04:00.
  • description — a constant label, "Form DRSLTR - Correspondence Related to Draft Registration Statement".
  • linkToFilingDetails — URL of the primary HTML document on www.sec.gov.
  • linkToTxt — URL of the EDGAR full-submission .txt bundle (the SGML concatenation of all documents).
  • linkToHtml — URL of the EDGAR filing index page for the accession.
  • linkToXbrl — empty string, since correspondence carries no XBRL.
  • documentFormatFiles — array describing every file in the original EDGAR submission, including those excluded from the ZIP (notably graphics).
  • entities — array of filer / subject-company / issuer entity objects parsed from the EDGAR submission header.
  • seriesAndClassesContractsInformation — empty array on DRSLTR (used only by investment-company filings).
  • dataFiles — empty array on DRSLTR (no machine-readable data exhibits).
  • id — internal 32-character hex identifier for the record.

The documentFormatFiles[] array contains one element per piece of the original submission. Each element carries sequence (a string sequence number assigned by EDGAR; the wrapper full-submission row uses a single space " "), documentUrl (a direct www.sec.gov URL), type (the EDGAR document-type tag — DRSLTR for the letter, GRAPHIC for embedded images, blank for the wrapper .txt), size (byte count as a string), and an optional description such as "GRAPHIC" or "Complete submission text file". The canonical pattern is one DRSLTR HTML at sequence 1, zero or more GRAPHIC rows at higher sequences, and a final wrapper-.txt row whose URL ends in <accession>.txt.

The entities[] array carries one element per filer or subject company in the EDGAR header. Each element typically contains companyName with a parenthetical role suffix ("Icon Energy Corp (Filer)", "... (Subject)", "... (Issuer)"), cik as a string without leading zeros, irsNo (occasionally the placeholder "000000000" when an EIN has not been issued — common for foreign private issuers), stateOfIncorporation as a two-letter code, fiscalYearEnd as MMDD (sometimes absent for foreign filers), sic combining the four-digit SIC code with its industry description (e.g. "6411 Insurance Agents, Brokers & Service"), type repeating the form code, and tickers, an array of ticker symbols associated with the CIK that may be empty for pre-IPO filers, single, or multi-class (e.g. ["JTAI","OXACW","OXAC","JTAIZ","JTAIW","OXACU"]).

Correspondence document (filename1.htm)

The primary letter file is wrapped in EDGAR's standard SGML document envelope. The opening lines declare the document type, sequence, and original filename, and the body is bracketed by <TEXT> markers:

1 <DOCUMENT>
2 <TYPE>DRSLTR
3 <SEQUENCE>1
4 <FILENAME>filename1.htm
5 <TEXT>
6 <html>
7 ...letter body...
8 </html>
9 </TEXT>
10 </DOCUMENT>

<TYPE> is always DRSLTR and <SEQUENCE> is always 1 for the primary letter. The HTML body is heterogeneously authored — common authoring toolchains include Broadridge PROfile, Workiva Wdesk, and hand-rolled legal-printer templates such as those produced by EdgarAgents — but the semantic content is consistent. A typical letter contains, in order:

  1. Outside-counsel letterhead, often referencing a logo image that is not packaged with the record.
  2. A delivery legend such as "VIA EDGAR" or "CONFIDENTIAL".
  3. The date line.
  4. The addressee block — typically U.S. Securities and Exchange Commission, Division of Corporation Finance, 100 F Street, N.E., Washington, D.C. 20549 — followed by named staff reviewers and an Office designation.
  5. An attention line directed to the staff accountant or staff attorney.
  6. A Re: subject line naming the issuer, the underlying form (most frequently S-1, F-1, or S-3), and the CIK.
  7. The salutation.
  8. The substantive body, which takes one of several recurring shapes: a short cover note announcing the confidential nonpublic submission of an initial or amended draft registration statement and certifying Emerging Growth Company status under Section 2(a)(19) of the Securities Act; a longer point-by-point response that quotes each numbered staff comment in italics or block-indent and follows it with the company's response, often accompanied by marked text excerpts proposed for inclusion in a subsequent amendment; a request that the staff direct further communications to named counsel; or a representation regarding the public-filing timing window (commonly 48 hours or two business days before requested effectiveness, per the JOBS Act and Division of Corporation Finance practice).
  9. A signature block introduced by /s/, the printed name of the signing attorney, and the law-firm affiliation.
  10. A cc: list naming issuer officers, additional counsel, and accounting advisers.

Additional attachments

Some accessions include further textual attachments alongside filename1.htm when the filer transmits multiple letters in a single submission, supplemental memoranda, or exhibits to the response (such as proposed disclosure mark-ups). These appear as additional HTML or — less commonly — TXT or PDF files in the same folder, with names matching the original EDGAR <FILENAME> tags. The metadata's documentFormatFiles[] array enumerates all of them, and the sequence field indicates the order in which EDGAR received them.

Included content

A record packages the structured metadata JSON plus every textual document originally transmitted in the EDGAR submission: the primary DRSLTR letter and any additional textual attachments. The metadata.json exposes the full EDGAR header (form type, accession, filing timestamp, filer entities, issuer identifiers, ticker symbols, SIC classification, state of incorporation, fiscal year end), the complete inventory of the original submission with direct www.sec.gov URLs, and the dataset's internal record identifier. The SGML wrapper around each letter preserves EDGAR's <TYPE>, <SEQUENCE>, and <FILENAME> markers so that the document boundary and original filename remain recoverable.

Excluded or separate content

Image files referenced by the HTML — typically law-firm logos, scanned signatures, and inline figures with names like image_001.jpg, formdrsltr_001.jpg, or issuer-specific filenames — are intentionally omitted. The HTML retains the <img src="..."> references, so a record opened in a browser will display broken image placeholders, but the corresponding documentUrl entries in metadata.json still point to the live copies on www.sec.gov for callers who wish to retrieve them. The full-submission .txt wrapper that EDGAR produces is not packaged as a file inside the folder; only its URL is recorded in documentFormatFiles[] and linkToTxt. The DRSLTR record also does not contain the underlying registration statement itself (the DRS / DRS/A filing) — that is a separate accession in a separate dataset; only the correspondence about the draft is in scope here. Series-and-classes/contracts information and machine-readable data files are absent because correspondence does not carry such content.

Changes in required content or structure over time

From inception in October 2012 through mid-2017, the form's effective use was restricted to EGC issuers; substantive content reflected that scope, with virtually every cover letter containing an explicit EGC representation and an undertaking to publicly file at least 21 days before any roadshow. Effective July 10, 2017, the Division of Corporation Finance expanded the availability of nonpublic review to any issuer's initial registration statement and to certain follow-on offerings within twelve months of an IPO, and the FAST Act technical corrections in 2018 codified parts of that practice. After mid-2017 the substantive content of DRSLTR letters broadened accordingly: many letters from larger or seasoned issuers no longer include an EGC representation, and the confidential-treatment basis cited shifts from "Section 106 of the JOBS Act" to the Division's expanded policy. The public-filing timing commitment cited in cover letters also shortened, with the JOBS Act's original 21-day pre-roadshow window replaced for non-EGC issuers by the 48-hour / two-business-day pre-effectiveness window adopted in the Division's 2017 guidance.

The structural envelope of the filing — SGML document wrapper, metadata header, optional attachments — has remained stable since 2012. There are no Item-level disclosure requirements that have been added or removed because DRSLTR is correspondence rather than a structured form.

Changes in data format over time

DRSLTR submissions have always been transmitted through the modern EDGAR system and have always carried the SGML document wrapper around an HTML or, less commonly, plain-text body. Because the form was introduced in October 2012, well after EDGAR's transition away from ASCII-only filings, there is no early text-only era to describe; HTML has been the dominant body format for the entire life of the dataset. PDF appears occasionally as an attachment when filers submit scanned or graphics-heavy supplemental materials, and TXT appears as an alternative body format in a small minority of older or simpler letters. XBRL is not applicable to this form. Beyond minor evolution in authoring toolchains — early letters often produced by simple word-processor HTML exports, more recent ones by Broadridge PROfile, Workiva Wdesk, and other professional-grade platforms — there has been no material format transition.

Interpretation notes

Several nuances matter for downstream use:

  • Unstructured prose. DRSLTR is correspondence, not a regulated disclosure form, so its content is unstructured prose. Extraction of staff comments and issuer responses depends on parsing recurring textual patterns — italicized or block-indented staff comments followed by company responses, numbered comment lists, "Re:" lines naming the underlying form — rather than on reading defined Items.
  • Toolchain heterogeneity. Because the body HTML is produced by a wide range of authoring tools, layout tags vary considerably; reliable text extraction should normalize across toolchains rather than assume a single template.
  • Resolving the primary document. The primary letter's filename is conventionally filename1.htm because EDGAR's submission template defaults to that name for DRSLTR letters, but this is a convention rather than a guarantee. Applications should resolve the primary document by reading the first documentFormatFiles[] entry in metadata.json to handle edge cases.
  • Dangling image references. Embedded <img> references will dangle inside the packaged HTML; rendering the letter with images intact requires fetching the referenced files from the URLs in documentFormatFiles[].
  • Multiple entity roles. The entities[] array can contain multiple roles — a single submission may list a filer, a subject issuer, and a parent entity — and the role suffix in companyName ((Filer), (Subject), (Issuer)) is the canonical way to disambiguate them.
  • Acceptance vs. publication timing. DRSLTR records become public on EDGAR only when the underlying registration statement is publicly filed, so the filedAt timestamp captured in metadata reflects the original confidential-submission acceptance, not the date the record became publicly accessible.
  • Letter size variability. Individual letters vary widely in size and density, from one-page cover notes of a few kilobytes to multi-section response letters of one hundred kilobytes or more of HTML, with longer letters typically corresponding to point-by-point responses to staff comment letters.

Who Files or Publishes This Dataset, and When

Form DRSLTR is an EDGAR correspondence submission type, not a registration form. The filer of record is the issuer undergoing confidential, non-public staff review of a draft registration statement (DRS). In practice, the substantive document is drafted and signed by the issuer's outside securities counsel on counsel's letterhead, but the EDGAR submission is made through the issuer's filer credentials.

Eligible filers fall into two regimes:

Typical DRSLTR filers therefore include domestic operating companies on Form S-1, foreign private issuers on Form F-1, Form 20-F, or Form 40-F, Form 10 registrants, recently public follow-on issuers within the twelve-month window, and certain business-combination registrants whose draft submissions are accepted for confidential review. Auditors, valuation specialists, and selling-shareholder counsel may contribute attached analyses, but they do not file DRSLTR records under their own names.

When the record is created or required

DRSLTR is event-driven, with no statutory deadline for any individual submission. A record is created when the issuer needs to communicate in writing with the staff during the confidential review track. The principal triggers are:

  • a written response to a staff comment letter on a previously submitted DRS or DRS/A, typically paired with a DRS/A amendment containing the revised filing
  • issuer-initiated requests for pre-submission conferences, accounting accommodations, financial-statement waivers, or confidential treatment
  • transmittal of supplemental analyses requested by the staff (cheap-stock and stock-compensation valuations, segment determinations, non-GAAP reconciliations, MD&A backup)
  • status, withdrawal, or scheduling communications tied to the confidential submission

The dominant pattern is comment-driven: an issuer submits a DRS, the staff issues comments (roughly 30 days for a first review, less for subsequent rounds), and the issuer responds with a DRSLTR plus a DRS/A. Each round generates a new DRSLTR record; DRSLTR submissions are not themselves amended.

Although the individual DRSLTR has no clock, the underlying confidential dossier does. Under Section 6(e) (as amended by FAST Act Section 71003), an EGC must publicly file the registration statement and all prior draft submissions and related correspondence at least 15 days before commencing a road show or, absent a road show, before requested effectiveness. The 2017 staff procedure imposes the same 15-day public-filing requirement before a road show or effectiveness for non-EGC IPOs and follow-ons, and at least 15 days before effectiveness for Forms 10, 20-F, and 40-F. Confidential DRS, DRS/A, and DRSLTR submissions become publicly available on EDGAR at that public-filing point. Staff comment letters themselves are released no earlier than 20 business days after effectiveness, consistent with the staff's general release policy.

Important distinctions

  • DRSLTR vs. CORRESP. Both are correspondence submission types under the EDGAR Filer Manual. DRSLTR carries correspondence during the confidential draft phase; CORRESP carries correspondence on publicly filed registration statements and Exchange Act reports. Once an issuer transitions to public filing, subsequent comment responses are filed as CORRESP.
  • DRSLTR vs. DRS / DRS/A. DRS and DRS/A are the draft registration statement and its amendments. DRSLTR is the separate correspondence track that accompanies them; it does not contain the prospectus.
  • Issuer as filer, counsel as author. The issuer is always the EDGAR filer of record. Outside counsel writes and signs the letter; auditors may attach schedules but do not file in their own name.
  • Loss of EGC status. An issuer loses EGC status on the earliest of (i) the last day of the fiscal year in which it exceeds the indexed revenue threshold, (ii) the last day of the fiscal year following the fifth anniversary of its IPO, (iii) issuance of more than $1 billion in non-convertible debt over three years, or (iv) qualification as a large accelerated filer. Such an issuer can generally still use the expanded 2017 non-public review track.
  • Withdrawn or never-public submissions. If a confidential submission is withdrawn and never publicly filed, the associated DRSLTR records typically remain non-public. The publicly visible DRSLTR population is thus biased toward issuers that progressed to a public registration.
  • No pre-2012 records. DRSLTR depends on the JOBS Act regime and has no pre-EDGAR analog; the earliest records date to 2012. The narrower pre-JOBS Act FPI confidential-review accommodation was paper-based and was not assigned an EDGAR submission type.

How This Dataset Differs From Similar Datasets or Filings

The Form DRSLTR dataset sits inside a tight cluster of pre-IPO and registration-review records on EDGAR. The most useful comparisons are with DRS/DRS/A, UPLOAD, CORRESP, the eventual public registration statements (S-1, F-1, S-11, Form 10/10-12B), and aggregated public comment-letter packages.

DRS and DRS/A — the draft registration statements themselves

DRS and DRS/A are the confidential draft registration statements (and amendments) submitted for non-public Division of Corporation Finance review. They contain the substantive disclosure: prospectus, financials, risk factors, MD&A, exhibits, signatures. DRSLTR is the cover-and-response correspondence that travels alongside these submissions. DRS is the disclosure document; DRSLTR is the issuer's narrative letter explaining how staff comments were addressed, what changed in a given amendment, and any supplemental analyses or accounting/legal positions. Researchers studying what was disclosed pull DRS; researchers studying the negotiation and revision history pull DRSLTR.

UPLOAD — SEC staff comment letters on draft registration statements

UPLOAD is the staff side of the same conversation: letters issued by Division staff raising comments on a draft registration statement. DRSLTR is the issuer's reply. The two should be read in pairs and typically iterate over several rounds until staff has no further comments. Filer population differs sharply: UPLOAD is authored and filed by SEC staff; DRSLTR is submitted by the issuer (usually through outside counsel). Both are held confidentially during review and become publicly visible on EDGAR only after the registration statement is publicly filed and the review concludes. Reconstructing the full back-and-forth requires joining both datasets on CIK and time order.

CORRESP — general correspondence on public filings

CORRESP and DRSLTR are functionally analogous issuer-to-staff response letters; they differ on confidentiality and timing. DRSLTR is used during the non-public draft review window; CORRESP is used once the filing is public. A confidential filer typically generates DRSLTRs during the draft phase, then switches to CORRESP once the registration statement is publicly filed and amendments draw further staff comments. Reconstructing a complete comment-and-response history of a single offering therefore usually requires splicing DRSLTR and CORRESP.

S-1, F-1, S-11, and Form 10 / 10-12B — the eventually-public registration statements

S-1 (domestic), F-1 (foreign private issuer), S-11 (real estate), and Form 10/10-12B (Exchange Act class registration) are the public registration statements that confidential DRS submissions convert into at the public phase. The disclosure text is often nearly identical to the final DRS/A, but filed under a public form type. DRSLTR sits on the upstream, non-public side of that boundary; these public forms sit on the downstream side. DRSLTR does not contain registration disclosures themselves — it contains correspondence about how those disclosures were drafted and revised.

Public comment-letter packages released after effectiveness

Once a registration statement becomes effective (or is abandoned for more than 180 days), Division staff publicly releases the full UPLOAD plus CORRESP/DRSLTR package on EDGAR, generally about 20 business days after the review closes. Aggregated comment-letter products bundle UPLOAD, CORRESP, and DRSLTR into a single corpus. The DRSLTR dataset is narrower: only issuer-authored response letters tied to the confidential draft phase, with no staff letters and no post-public correspondence mixed in.

Key differences at a glance

  • Authorship: DRSLTR is issuer-authored; UPLOAD is staff-authored.
  • Confidentiality: DRSLTR and UPLOAD originate in the non-public draft window; CORRESP applies to public filings.
  • Document type: DRSLTR is correspondence; DRS, S-1, F-1, S-11, and Form 10/10-12B are disclosure documents.
  • Timing: DRSLTR is generated during draft review and becomes publicly visible on EDGAR only after the review concludes (typically post-effectiveness or 180-day abandonment).
  • Scope: DRSLTR isolates the issuer-response slice; merged comment-letter corpora cover both sides and both phases.

Boundary summary

What makes DRSLTR distinct is the intersection of four constraints: confidential draft-review origin (under JOBS Act Section 106, expanded by FAST Act Section 71003 and the Division's December 2017 broadening to most IPOs and certain follow-ons within twelve months); issuer-side authorship; correspondence content rather than registration disclosure; and a per-round trigger keyed to specific staff comments, with delayed public availability. UPLOAD captures the staff voice but not the issuer's. CORRESP captures issuer correspondence but only post-public. DRS captures the draft disclosure document but not the dialogue. S-1/F-1/S-11/10-12B capture the final public registration but not the iterative drafting record. Only DRSLTR isolates issuer response letters generated during the confidential review phase.

Who Uses This Dataset

DRSLTR filings are the issuer-side responses to SEC staff comments during confidential pre-IPO review of draft registration statements. They expose how registrants and counsel argued through accounting policies, segment reporting, non-GAAP measures, executive compensation, and risk-factor disclosures. Three parts of each record carry most of the value: the metadata.json filer entity block (CIK, issuer name), the HTML or PDF letter body with numbered response paragraphs, and accession metadata (accession number, filed date, signing counsel).

Securities and capital-markets lawyers

IPO and disclosure counsel use the corpus as a working library of how peer issuers answered staff comments. They pull comparables by SIC code, counsel signature, and filed date, then mine the response narrative for how registrants framed non-GAAP objections, segment aggregation, revenue-recognition timing, or pro forma treatments. Output: faster, better-calibrated draft responses and pre-empted comments in the original S-1.

Equity capital markets bankers

ECM and syndicate desks track which issuers are in confidential review and how far they have progressed. Accession sequences and filed dates against a single CIK reveal the cadence of staff rounds; response content signals which issues remain open. Used for IPO pipeline maps, deal-readiness assessments, and launch-window timing.

Pre-IPO and new-issue research analysts

Analysts surface issues raised in pre-public review that the final S-1 may smooth over. They extract response paragraphs on KPI definitions, customer concentration, related-party transactions, and going concern matters to build pre-deal notes, valuation cautions, and management Q&A for roadshows.

In-house compliance and disclosure officers

Issuer compliance teams and disclosure consultants run text searches across letter bodies to identify recurring staff focus areas: segment reporting, revenue recognition, lease accounting, cybersecurity, climate risk, MD&A trends. Used for internal training, drafting checklists, and pre-review readiness scoring.

Technical accounting and SEC services groups at audit firms

Accounting specialists treat DRSLTR responses as a primary source on how registrants resolve technical questions before effectiveness. They focus on responses citing specific ASC sections, SAB topics, and non-GAAP guidance, plus supporting tables. Used for technical alerts, IPO readiness checklists, and benchmarking accounting conclusions.

Academic researchers in accounting, finance, and law

Researchers link letters to matching DRS filings and later S-1 amendments via CIK and accession number, then apply NLP to comment themes, response length, and revision intensity. Supports empirical studies on disclosure quality, IPO outcomes, and regulatory behavior.

Fintech and disclosure-data vendors

Product teams ingest the metadata for issuer identity and timeline normalization, then parse letter HTML to extract numbered comments and response paragraphs. Outputs include IPO pipeline trackers, topic classifiers (revenue, segments, going concern, compensation), and entity-resolution pipelines linking DRSLTR letters to subsequent S-1, 424B, and 10-K filings.

Activist and short-biased research desks

Short-side analysts mine the response narrative for disclosure weaknesses the issuer had to defend in private: aggressive revenue recognition, related-party concentration, going-concern flags, weak internal controls. Used for short-thesis memos, governance-risk flags on newly public issuers, and proxy-season targets.

Corporate general counsel and securities-law teams

Issuers preparing their own IPO or follow-on registration assemble comparable correspondence by counsel firm, SIC code, and date range to scope expected staff comments and instruct outside counsel.

Investor-relations advisors to pre-IPO issuers

IR consultants coach management on the disclosure questions the SEC most often raises in their sector and the language registrants use to resolve them. They extract response excerpts on KPIs, non-GAAP measures, and risk-factor framing for pre-roadshow messaging and draft Q&A.

NLP and RAG developers in financial AI

Teams building securities-law assistants use DRSLTR letters as high-signal training and retrieval material because each response pairs a specific staff question with a substantive issuer answer. Metadata JSON provides clean entity grounding; HTML supports chunking and embedding for comment-letter classifiers and precedent-response retrieval.

Specific Use Cases

The DRSLTR corpus is most useful when the workflow depends on issuer-side response language, draft-review timing, or counsel-firm patterns. Each of the following uses ties directly to the metadata, the numbered response narrative, or the accession-level anchors of the records.

Building a precedent library of staff-comment responses by topic

IPO counsel and audit-firm technical accounting groups query letter bodies for recurring topics — non-GAAP reconciliations, segment aggregation under ASC 280, revenue-recognition timing under ASC 606, going-concern language, cybersecurity disclosure — and pull the surrounding response paragraphs as drafting precedent. The entities[] block (CIK, SIC, state of incorporation) supports peer filtering by industry, and the SGML-wrapped HTML preserves numbered response structure for clean chunking. Output: a searchable response-precedent library used to pre-empt staff comments in initial S-1 drafts.

Mapping the IPO confidential-review pipeline

ECM desks and pipeline-tracking vendors group DRSLTR accessions by CIK and order them by filedAt to reconstruct the cadence of confidential review rounds for each issuer. Round count, gap between letters, and signing counsel from the letter body indicate deal readiness and likely launch window. Output: a pre-IPO pipeline tracker that flags issuers nearing the public-filing trigger and projects launch timing.

Reconstructing full staff-issuer dialogues by joining UPLOAD and DRSLTR

Researchers and disclosure-data vendors pair DRSLTR records with UPLOAD staff comment letters on CIK and chronological order to assemble complete comment-and-response threads, then splice in CORRESP letters once the registration becomes public. The accessionNo, filedAt, and filer CIK in metadata.json supply the join keys. Output: round-by-round dialogue datasets used in NLP studies of disclosure negotiation, regulator behavior, and revision intensity.

Benchmarking outside-counsel performance and market share

Lawyers, ECM bankers, and corporate general counsel selecting IPO counsel parse the signature block and cc: list of each letter — together with the issuer SIC and state of incorporation in metadata — to tally which firms handle which sectors, how many rounds their deals require, and how their response style differs across topics. Output: counsel league tables, sector-specific firm shortlists, and informed pitches against named competitors.

Surfacing pre-public disclosure issues for short-side and pre-deal research

Short-biased and pre-IPO analysts mine response paragraphs for issues the issuer had to defend privately — aggressive revenue recognition, customer concentration, related-party transactions, internal-control weaknesses, going-concern triggers — and link them to the eventual public S-1 to identify language that was softened or removed. Output: short-thesis memos, governance-risk flags on newly public issuers, and pre-deal research notes.

Training and grounding securities-law AI assistants

NLP and RAG developers use DRSLTR records as a high-signal corpus because each response paragraph pairs a specific staff question with a substantive issuer answer. The metadata.json entity block provides clean grounding (CIK, issuer name, SIC, tickers); the SGML-wrapped HTML supports deterministic document-boundary detection and chunking. Output: comment-letter topic classifiers, precedent-response retrieval indexes, and grounded drafting assistants for IPO counsel.

Dataset Access

The Form DRSLTR Files Dataset is available through three access methods: a metadata index endpoint, a full dataset archive, and individual monthly container downloads.

Dataset Index JSON API: https://api.sec-api.io/datasets/form-drsltr-files.json

This endpoint returns dataset-level metadata (name, description, last updated timestamp, earliest sample date 2012-10-01, total records and total size, covered form types, container format, and contained file types) along with the full list of monthly container files. Each container entry includes its key, size, record count, last updated timestamp, and direct download URL. Use this index to monitor which containers were updated in the most recent refresh run and to decide which containers to download on a daily basis. This endpoint does not require an API key.

Example response:

Example
1 {
2 "datasetId": "1f13365b-9ae0-6947-89ff-baa6e2db28bd",
3 "datasetDownloadUrl": "https://api.sec-api.io/datasets/form-drsltr-files.zip",
4 "name": "Form DRSLTR Files Dataset",
5 "updatedAt": "2026-04-25T03:02:39.982Z",
6 "earliestSampleDate": "2012-10-01",
7 "totalRecords": 5884,
8 "totalSize": 158824389,
9 "formTypes": ["DRSLTR"],
10 "containerFormat": "ZIP",
11 "fileTypes": ["HTML", "JSON", "TXT", "PDF"],
12 "containers": [
13 {
14 "downloadUrl": "https://api.sec-api.io/datasets/form-drsltr-files/2026/2026-04.zip",
15 "key": "2026/2026-04.zip",
16 "size": 1342118,
17 "records": 47,
18 "updatedAt": "2026-04-25T03:02:39.982Z"
19 }
20 ]
21 }

Download Entire Dataset: https://api.sec-api.io/datasets/form-drsltr-files.zip?token=YOUR_API_KEY

Downloads the complete Form DRSLTR Files Dataset as a single ZIP archive containing all monthly containers from the earliest sample date forward. This endpoint requires a valid SEC API key.

Download Single Container: https://api.sec-api.io/datasets/form-drsltr-files/2025/2025-08.zip?token=YOUR_API_KEY

Downloads one monthly container ZIP holding all DRSLTR filings for that month, including the original HTML, TXT, and PDF source files plus a JSON metadata file per filing. Replace the year and month segments with any container key returned by the index JSON. This endpoint requires a valid SEC API key.

Frequently Asked Questions

What form does this dataset cover?

The dataset covers EDGAR submission type DRSLTR — issuer correspondence transmitted to the SEC's Division of Corporation Finance during the confidential, non-public review of a draft registration statement (filed under EDGAR types DRS or DRS/A). It does not include the underlying draft registration statement itself, which is a separate EDGAR submission type.

What does one record in this dataset represent?

One record represents a single EDGAR DRSLTR submission — one accession number — and is materialized as a folder named with the eighteen-digit unpunctuated accession number. The folder contains a metadata.json envelope plus one or more SGML-wrapped HTML letters (and occasionally TXT or PDF attachments) that the issuer transmitted under that accession.

Who is required to file Form DRSLTR?

The filer of record is always the issuer, although the letter is almost always drafted and signed by outside securities counsel. Eligible filers are emerging growth companies under Section 6(e) of the Securities Act (added by JOBS Act Section 106) and, since the Division of Corporation Finance's expanded procedures effective July 10, 2017, any issuer pursuing an initial registration statement (S-1, F-1, S-11, Form 10, 20-F, or 40-F) or a follow-on offering within twelve months of an IPO or initial Exchange Act registration.

What time period does the dataset cover?

The dataset's earliest sample date is 2012-10-01, which is when Form DRSLTR was introduced into EDGAR alongside Form DRS to operationalize the JOBS Act confidential-submission regime. There are no pre-2012 records because the form has no pre-EDGAR analog. Coverage runs forward from October 2012 to the most recent monthly container.

What file format is the dataset distributed in?

Records are delivered as monthly ZIP containers. Inside each container, every record is a folder named with the unpunctuated accession number, holding a metadata.json plus the textual correspondence files. File types found across the dataset are HTML, JSON, TXT, and PDF; embedded images referenced from the letter HTML are intentionally excluded.

How does this dataset differ from CORRESP and UPLOAD?

DRSLTR is issuer-authored correspondence during the confidential draft-review phase. CORRESP is functionally analogous issuer correspondence but only on publicly filed registration statements and Exchange Act reports — it kicks in once a filing goes public. UPLOAD is the staff side of the conversation: comment letters issued by Division of Corporation Finance staff. Reconstructing a complete comment-and-response history typically requires joining DRSLTR with UPLOAD on CIK and time order, then splicing in CORRESP for the post-public phase.

Why are some DRSLTR submissions not visible on EDGAR?

DRSLTR submissions are nonpublic at acceptance and become publicly visible on EDGAR only after the issuer publicly files the registration statement. If a confidential submission is withdrawn and never publicly filed, the associated DRSLTR records typically remain non-public, so the visible DRSLTR population is biased toward issuers that progressed to a public registration.