@pipeworx/reginfo

Connect: https://gateway.pipeworx.io/reginfo/mcp · Install: one-click buttons

Tools: 4

The Unified Agenda of Federal Regulatory and Deregulatory Actions — every RIN (Regulation Identifier Number) a federal agency has told OIRA it intends to propose, finalize, or has completed, published biannually since Fall 1995. This is the LEADING record ahead of ecfr/federal-register’s trailing one: an agency’s “planning to propose X” shows up here, edition by edition, often years before the same rule actually publishes or takes effect.

Tools

• reg_agenda_search(agency?, keyword?, stage?, edition?, limit?) — search RINs by agency, free-text keyword, and/or rulemaking stage. Defaults to the newest edition on file.
• reg_rin(rin) — full detail for one RIN across EVERY edition it has appeared in, showing how its stage/status changed over time.
• reg_agency_agenda(agency, edition?, limit?) — one agency’s full regulatory agenda for an edition, grouped by rule_stage.
• reg_agenda_coverage() — which editions are currently loaded (oldest/newest, total rows).

Auth

Keyless — no operator key needed. Backed by reginfo_rins, refreshed monthly by workers/data-pipeline (dataset reginfo).

Data sources

• https://www.reginfo.gov/public/do/XMLReportList — the Unified Agenda report index.
• https://www.reginfo.gov/public/do/XMLViewFileAction?f=REGINFO_RIN_DATA_202504.xml — one biannual edition’s full XML (~16.5MB, ~3,800 RIN_INFO records). Editions run REGINFO_RIN_DATA_199510.xml .. present, April(04)/October(10) each year.

A missing edition answers HTTP 200, not 404. reginfo.gov serves a ~50KB HTML “no such report” page (Content-Type: text/html) for an edition that was never published — a real edition answers Content-Type: application/xml. The ingest HEAD-probes content-type before ever GETting a file, so a not-yet- published edition (or a genuine off-cycle gap in the pre-2000 grid) costs a cheap HEAD request rather than leaking an HTML page into the data.

One row PER (rin, publication_id), not per RIN. A RIN’s rule_stage changes edition to edition (“Proposed Rule Stage” → “Final Rule Stage” → “Completed Actions”) — that history is the whole point of this dataset, so a newer edition never overwrites an older one. reg_rin returns every edition a RIN has appeared in, newest first.

Coverage grows in gradually, not all at once. The backfill drains in edition by edition rather than loading all ~60 editions back to 1995 in one pass; reg_agenda_coverage reports exactly which editions are loaded right now, so a caller can tell “not backfilled yet” apart from “genuinely never published”.

Agency matching covers name, acronym, AND code, and matches on either the direct agency or its PARENT — asking for “Interior” also surfaces its sub-agencies’ (e.g. Fish & Wildlife) rules, since the source records both levels per RIN.

The abstract is HTML wrapped in CDATA in the source XML. The ingest strips tags and decodes entities before writing the abstract column, so callers only ever see plain text.

Tools

• reg_agenda_search — Search the Unified Agenda of federal regulatory actions — every rule an agency has told OIRA it plans to propose, finalize, or has completed, since Fall 1995. Answers “what rules is EPA planning”, “re
• reg_rin — Full detail for one Regulation Identifier Number (RIN), across EVERY Unified Agenda edition it has appeared in — proving the rule’s history rather than just its latest status. Returns the abstract, ag
• reg_agency_agenda — One agency’s full regulatory agenda for a given (or the newest) edition — every RIN it has open, grouped by rule_stage. Answers “what is EPA’s regulatory agenda”, “how many rules does DOL have in the
• reg_agenda_coverage — Reports which Unified Agenda editions are currently loaded — the oldest and newest edition on file, total edition count, and total RIN-edition rows. Use to gauge coverage before relying on this for a