Federal Register Scraper — New Rules & Comment Deadlines

Search the US Federal Register by keyword, agency, document type, CFR title and part or topic and get one clean row per document: title, type, abstract, agencies, publication date, effective date, comment deadline, docket IDs, RINs, CFR references, page range, significant flag, plus the link to the document, its PDF and the regulations.gov comment link. Final rules, proposed rules, notices and presidential documents, back to 1994. $2.00 per 1,000 documents, no login, no API key.

Run it again — or put it on a daily schedule — and every row also tells you whether the document is new since the last run (isNew). Set What to return to Only new and you pay only for documents you have not seen before. Switch on Comment period still open and you get the open dockets with their deadlines.

What you get

Field	What it is
documentNumber, url, pdfUrl	the Federal Register document number (for example 2026-21155), the link to its page on federalregister.gov and to the PDF of the printed edition on govinfo.gov
title, abstract	the title as published, and the abstract cut at a word near 500 characters (null when the document has none)
documentType, subtype, action	Rule, Proposed Rule, Notice or Presidential Document; for presidential documents the kind (Proclamation, Memorandum…); the action line the agency printed (Final rule., Notice of proposed rulemaking., 60-Day notice.)
agencies, agencySlugs	the agencies as federalregister.gov names them (a department and its sub-agency appear as two entries) and their slugs, which you can pass back in Agencies
publicationDate, effectiveDate	when the Federal Register published it and, where the document states one, when it takes effect
commentsCloseDate, commentPeriodOpen, commentUrl	the date comments are due; true / false whether that date is today or later in US Eastern time on the day of the run (null when there is no deadline); the regulations.gov link for commenting, where there is one
docketIds, regulationIdNumbers	the docket numbers and the RINs (Regulation Identifier Numbers), each as the agency printed them
cfrReferences, topics	the CFR parts the document affects (40 CFR 80, or 5 CFR Chapter XIV when only a chapter is given) and its CFR indexing topics
citation, startPage, endPage, pageCount	the Federal Register citation (91 FR 65699) and the page range in the printed edition
significant	true / false whether the Federal Register marks the rulemaking as significant under Executive Order 12866; null when it has no flag
isNew, previousRunAt	true / false whether you had not seen this document in your previous run of the same search, and when that run was (null on the first run)
observedAt	when the row was collected

Run it again and again

• On a schedule — save your input as a task and add a Schedule: the same search every morning. Send each run to Google Sheets, Slack or email with an integration on the task.
• Watch for changes — Set What to return to Only documents that are new since the last run.
• In bulk — Combine Keywords, Agencies, Document types and a CFR title in one run; each keyword line is its own search and the results come back as one list, newest first.

Use it for

• Regulatory monitoring — a daily schedule with your keywords or agencies, Only new, sent to Slack, email or your compliance tool
• Comment deadlines — Comment period still open with your agencies or CFR part: every open docket, its deadline and the regulations.gov link, soonest to act on
• Policy and lobbying research — Proposed rules and Significant rules only for a sector, with RINs and dockets to follow a rule from proposal to final
• Compliance and legal teams — everything that touches 40 CFR 60 or 21 CFR since a date, with CFR references and effective dates

Input

{
  "keywords": ["cybersecurity", "\"data breach\""],
  "agencies": ["Securities and Exchange Commission", "EPA"],
  "documentTypes": ["rule", "proposed-rule"],
  "commentPeriodOpen": true,
  "maxResults": 200,
  "output": "new"
}

• All filters are combined with AND; the entries inside one list are combined with OR.
• keywords — searched in the full text of the document. Several words on one line must all appear; put a phrase in quotes. Every line is a separate search (up to 10) and the results are merged without duplicates.
• agencies — the agency's name, abbreviation or slug (Environmental Protection Agency, EPA, environmental-protection-agency). The Actor looks the name up in the Federal Register's agency list and stops before charging anything if it finds no match or an abbreviation that fits two agencies (it then lists their slugs).
• documentTypes — rule, proposed-rule, notice, presidential-document. Empty means all four.
• significantOnly — only rulemakings marked significant under Executive Order 12866.
• cfrTitle, cfrPart — documents that affect that CFR title (1–50) and, if given, part (60) or part range (50-99). A part needs its title.
• topics — CFR indexing topic slugs as listed at federalregister.gov/topics, such as air-pollution-control.
• commentPeriodOpen — only documents whose comment deadline is today or later (US Eastern date). When you give neither Published since nor Published within, it looks back 730 days.
• publishedSince (YYYY-MM-DD) or publishedWithinDays (default 7) — the publication period, up to today.
• Results are always sorted newest first, and maxResults (default 200, up to 5,000) cuts the list from the old end. With Only new it is the number of new documents to return: the Actor reads down the list, past the documents you have already seen (up to 5,000 deep), until it has found that many. Documents below the cut-off are not marked as seen, so they come back as new in your next run.

"Is it new" is remembered per search — the same keywords, agencies, document types, significant flag, CFR title and part, topics and comment-period setting — in a named key-value store in your Apify account, so the second run of the same search already shows isNew. Changing Published within, Published since or Maximum documents keeps the same history.

Pricing

$2.00 per 1,000 documents, plus $0.01 per run. Error rows (a search the API rejects, the API not answering) are free. With Only new, documents you have already seen are skipped and not charged. The $0.01 run-start fee is the one charge every run pays, even one that returns nothing.

Notes

• Source: the public Federal Register API (federalregister.gov/api/v1) run by the Office of the Federal Register (National Archives) and the Government Publishing Office. Anonymous access, no key, nothing from behind a login. The site's robots.txt disallows only a few web pages (its search pages, /my/, /auth/), not /api/, and its About page says any person may reproduce or republish material from the Federal Register (1 CFR 2.6). Republishers may not use the official NARA or OFR logos or seals; this Actor uses none.
• Not an official legal edition. FederalRegister.gov is an unofficial informational resource: it does not provide legal notice to the public or judicial notice to the courts. Anyone relying on this data for legal research should check it against an official edition of the Federal Register (the PDF on govinfo.gov, linked in pdfUrl). This Actor is not affiliated with or endorsed by the Office of the Federal Register, the National Archives or the Government Publishing Office.
• How deep a search can go. The API documents that you can page through the first 2,000 results of a search and that you should use date ranges to get more. The Actor stays inside that: when you ask for more than 2,000 documents (Maximum documents above 2,000, or Only new, which reads up to 5,000 deep) it first counts the matches and halves the publication period until each part holds at most 2,000 documents, then reads the newest part first. A single day with more than 2,000 matching documents cannot be split further; you get 2,000 of them and a free error row saying the list may be incomplete. For scale: 436 documents were published on the four publication days 13–16 Oct 2026. Reading 2,500 documents from a 40-day period took 7 seconds in 7 requests on 16 Oct 2026.
• The API has no OR between search words (a check on 16 Oct 2026 gave the same count for A OR B as for A B), which is why each keyword line is run as its own search.
• abstract is empty on many notices: 103 of those 436 documents had none (meeting notices and filings mostly). Abstracts are cut at a word near 500 characters; the cut is marked with … (134 of the 333 abstracts in that read). A sentence of an abstract that carries an e-mail address, a phone number or a "please contact / directed to" line is dropped (2 of 333).
• No person fields. The Actor never asks for signers, authors or contact points, and agencies are organisations. Titles are passed on exactly as the Federal Register publishes them, and titles and abstracts can mention people or companies by name when the document does.
• commentsCloseDate was filled on 133 of those 436 documents, commentUrl on 96. 92 had both, 41 had a deadline without a regulations.gov link (27 of them information-collection, Paperwork Reduction Act or Privacy Act notices) and 4 had a link without a deadline. commentPeriodOpen is calculated at the time of the run from commentsCloseDate, so it is null when there is no deadline. A few documents carry deadlines years ahead (one final rule in that read: 25 Nov 2028); they are passed on as published.
• significant comes from the rulemaking's entry in the Unified Agenda: 5 documents were true, 29 false and 402 null in that read, and 34 of the 84 rules and proposed rules had a flag. Significant rules only therefore returns the flagged rules and leaves out the documents without a flag.
• docketIds are the agency's own labels and the format varies (Docket No. FDA-2022-N-2226, OMB Number 1105-0103, Release No. IC-36358). 312 of the 436 documents had at least one, and 53 had a RIN. cfrReferences was filled on 84 (up to 16 parts on one document). commentUrl is the regulations.gov link exactly as the API returns it (http://).
• A few agency entries on federalregister.gov have no slug (for example "Office of the Secretary" under a department); they appear in agencies but not in agencySlugs.
• If the API is rate limiting, the Actor waits and retries twice. Requests are spaced 0.7 seconds apart. If it still gets no answer it adds a free error row, and the run is marked failed when nothing at all could be read. A topic slug the API does not know is answered with an error row that names the field.

If you also watch public procurement, see EU Tenders Scraper (TED); for US vehicle safety recalls, see NHTSA Vehicle Recalls.

For AI agents

{"documentTypes": ["proposed-rule"], "commentPeriodOpen": true, "agencies": ["EPA"], "maxResults": 5} returns up to 5 open-for-comment EPA proposed rules, newest first, each with title, abstract, comment deadline, docket and the regulations.gov link. Rows have type: document or error. Fields that do not apply to a document are null (or an empty list), never missing. Schedule the same input daily with "output": "new" to get only what was published since the previous run.

**Something broke or looks wrong?** Open an issue on this Actor's Issues tab and I'll look at it. **Found it useful?** A short review on the Store page helps other people find it.