Federal Register Monitor — New Rules, Proposed Rules & Notices

Track newly published Federal Register documents by full-text keyword, agency and document type — final rules, proposed rules, notices and presidential documents. Official Federal Register API, with a change feed that remembers document numbers so nothing is missed or repeated.

What does this Actor do?

You give it keywords, agency slugs and document types. It queries the Federal Register's official API for each keyword separately, merges and de-duplicates the results by document number, sorts them newest-first, and emits them with a link to both the HTML and the PDF.

The interesting problem here is that Federal Register documents can appear before their publication date. A rule scheduled for Friday's issue is often available on Wednesday. Any monitor that tracks "what has been published since date X" will therefore either re-send those documents once their publication date arrives, or miss them entirely depending on which side of the comparison it lands.

This Actor keeps a list of document numbers it has already seen, alongside the run watermark, and uses that list — not the publication date alone — to decide what is new. Documents that surfaced early are flagged as new exactly once, on the run that first saw them.

Who is it for?

• Regulatory affairs and compliance teams who must know when a rule affecting their industry moves.
• Government affairs and policy analysts tracking a specific agency's rulemaking.
• Legal teams watching comment periods open on proposed rules.
• Trade associations and consultancies monitoring a sector across many agencies at once.
• AI agent builders who want a regulatory tool with a stable output schema.

Use cases

• Watch every SEC proposed rule mentioning a term and route it to a policy-review workflow.
• Monitor a set of agencies for any notice, as an early signal before formal rulemaking.
• Track a topic — artificial intelligence, PFAS, tariff — across all agencies at once.
• Feed newly published final rules into a compliance-obligations register.
• Give an agent a tool that answers "what has this agency published on this topic recently?"

What you get

One row per document. Any value the source does not publish is null — the field is always present, and the untouched source record stays in raw.

Field	Type	Description
documentNumber	string	Federal Register document number — the unique ID, and what the seen-list tracks
title	string	Document title
abstract	string	Federal Register abstract, where published
type	string	Document type as the source labels it, e.g. Rule, Proposed Rule, Notice
publicationDate	string	Publication date (YYYY-MM-DD)
agencies	array	Names of the issuing agencies
htmlUrl	string	Link to the document on federalregister.gov
pdfUrl	string	Link to the official PDF
commentsCloseOn	string | null	Public-comment deadline (YYYY-MM-DD), where published
docketIds	array | null	Federal Register docket IDs, where published
regulationIdNumbers	array | null	Regulation Identifier Numbers (RINs), where published
cfrReferences	array | null	CFR references formatted as title CFR part, where published
daysUntilCommentsClose	integer | null	Days until the public-comment deadline, calculated when the record is retrieved
raw	object	The complete original record from the API, unmodified
newSinceLastRun	boolean	true if this document number had not been seen on an earlier successful run
retrievedAt	string	ISO 8601 timestamp of when this row was fetched

Sample output record

{
  "documentNumber": "2026-17342",
  "title": "Safeguards for Automated Decision Systems in Consumer Finance",
  "abstract": "The Bureau proposes to amend Regulation B to address the use of automated decision systems…",
  "type": "Proposed Rule",
  "publicationDate": "2026-08-21",
  "agencies": ["Consumer Financial Protection Bureau"],
  "htmlUrl": "https://www.federalregister.gov/documents/2026/08/21/2026-17342/safeguards-for-automated-decision-systems",
  "pdfUrl": "https://www.govinfo.gov/content/pkg/FR-2026-08-21/pdf/2026-17342.pdf",
  "commentsCloseOn": "2026-10-20",
  "docketIds": ["CFPB-2026-0012"],
  "regulationIdNumbers": ["3170-AA99"],
  "cfrReferences": ["12 CFR 1002"],
  "daysUntilCommentsClose": 58,
  "raw": { "document_number": "2026-17342", "…": "…" },
  "newSinceLastRun": true,
  "retrievedAt": "2026-08-23T09:30:12.004Z"
}

How to use it

Watch one agency for rulemaking on a topic

{
  "keywords": ["artificial intelligence"],
  "agencies": ["securities-and-exchange-commission"],
  "documentTypes": ["RULE", "PRORULE"],
  "maxResults": 25
}

Everything from a set of agencies

Leave keywords empty to search the window without a term.

{
  "agencies": ["environmental-protection-agency", "food-and-drug-administration"],
  "maxResults": 200
}

Agency values are Federal Register slugs, not display names — the full list is at federalregister.gov/agencies.

Input parameters

Input	Type	Description
keywords	array	Full-text terms, each queried separately. Empty searches the window without a term
agencies	array	Federal Register agency slugs. Empty includes all agencies
documentTypes	array	RULE, PRORULE, NOTICE or PRESDOCU. Empty includes all types
publishedSince	string	YYYY-MM-DD start date. Overrides the previous successful run's watermark
maxResults	integer	Maximum records per run (1–1000). Default: 100
closingWithinDays	integer	Only emit documents whose public-comment deadline is from now through this many days away; documents without a deadline are excluded

Passing a document type outside the four allowed values fails the run with a clear error rather than silently returning nothing.

How monitoring works (first run and scheduling)

The first run returns documents from the last seven days unless publishedSince is supplied. To keep the watermark and the seen-document list across runs, schedule this Actor as a saved Task so runs share the same storage. A fresh unsaved run starts a fresh baseline.

The dataset contains the whole window, not only the new documents. Every run returns the documents matching your query in the window, and newSinceLastRun marks the ones this Actor had not seen before. Filter on that flag downstream if you only want to alert on genuinely new items.

Honest limitations

• Each run emits the window, not a strict delta. As above — use newSinceLastRun to separate new documents from ones you have already seen.
• Agencies must be given as slugs. securities-and-exchange-commission, not SEC and not Securities and Exchange Commission. An unrecognized slug simply returns nothing for that agency.
• Keywords are queried separately, then merged. Two keywords means two queries and a union of the results, not an AND across both terms.
• maxResults applies after the merge, so with several keywords the newest documents across all of them win and older matches on a single keyword may be cut.
• Full document text is not returned. You get the title, abstract and links; fetching and parsing the rule itself is a separate job.
• Some documents have no public-comment deadline or regulatory references. Their corresponding nullable fields are null; closingWithinDays excludes them.

Agent example

Ask an agent to run this Actor with keywords: ["artificial intelligence"] and send each resulting Rule or Proposed Rule to your policy-review workflow.

Reliability

Diagnostics are written to the RUN_SUMMARY key-value record — never into the dataset. A failed upstream request leaves the watermark and the seen-document list unchanged, so a transient outage cannot silently skip a window.

Every record carries retrievedAt and both htmlUrl and pdfUrl back to the official source, so you can check any row against the publisher yourself.

Maintained against the upstream API. If the Federal Register changes its schema or an endpoint moves, report it through the Issues tab and it gets fixed.

Integrations

Connect this Actor to Make, Zapier, n8n, Slack, Google Sheets, Airtable or any HTTP endpoint through Apify integrations. A daily Task can post new proposed rules into a policy channel, or fire a webhook so a summarization job starts the moment a rule publishes. Datasets export as JSON, CSV, Excel, XML, RSS or HTML.

API usage

curl -X POST "https://api.apify.com/v2/acts/oneshotventure~fedreg-watch/runs?token=YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"keywords": ["artificial intelligence"], "documentTypes": ["RULE", "PRORULE"], "maxResults": 25}'

Python, JavaScript, PHP and CLI clients are documented under Apify API clients.

Use with AI agents (MCP)

This Actor is callable from any MCP-compatible client — Claude, Cursor, VS Code or your own agent — through the Apify MCP server. The input schema is fully described and every record uses one stable, flat JSON shape, so an agent can call it and read the result without a parsing step. Because it reads the official Federal Register API rather than a rendered page, the answer an agent gets is the same one the source publishes — and htmlUrl gives the agent a document to fetch and summarize.

Frequently asked questions

Is there a free Federal Register API?

Yes — the Office of the Federal Register publishes an open API with no key required. This Actor is the scheduling, multi-keyword and change-feed layer over it, including the seen-document tracking that handles early-published documents.

How do I track new federal regulations by agency?

Put the agency's slug in agencies, save the Actor as a Task, and schedule it. Slugs are listed at federalregister.gov/agencies.

How do I monitor proposed rules in my industry?

Set documentTypes to ["PRORULE"] and add your industry terms to keywords. Add agency slugs to narrow further.

What is the difference between a rule and a proposed rule?

A proposed rule (PRORULE) is an agency's draft, published to invite public comment. A final rule (RULE) is the binding version published after that process. NOTICE covers announcements that are not rulemaking, and PRESDOCU covers presidential documents.

Why am I seeing documents I already received?

Each run returns the whole window, not only the delta. Filter on newSinceLastRun to get just the documents this Actor had not seen before.

Does it return the full text of the rule?

No. You get the title, abstract, type, agencies and links to the HTML and PDF. Retrieving and parsing the document itself is a separate step.

How far back can I go?

Set publishedSince to any date you like. Without it, the first run covers the last seven days and later scheduled runs continue from the watermark.

Related actors

• SAM.gov Contract Opportunities & Awards Feed — federal contract opportunities and awards.
• Grants.gov Monitor — new federal grant opportunities.
• US Product Recalls Monitor — CPSC and FDA enforcement actions.
• SEC EDGAR Filings Monitor — new company filings from EDGAR.

Disclaimer

This independent tool uses public U.S. government data and is not affiliated with the Office of the Federal Register.