Govinfo MCP Server

@pipeworx/govinfo

GovInfo.gov MCP — full text of US government publications.

Part of Pipeworx — an MCP gateway connecting AI agents to 1679+ live data sources.

Tools

• list_collections()
• search_packages(query, collections?, congress?, date_from?, date_to?, page_size?, offset_mark?)
  • collections takes one code or several, comma- or space-separated ("CHRG", "CHRG,CRPT"). Several codes search the union — "CHRG,CRPT" returns 9,239 hits for "Cooper", exactly the 5,359 + 3,880 the two collections return on their own.
  • page_size also answers to limit.
• get_package(package_id)
• list_granules(package_id, page_size?, offset_mark?) — page_size also answers to limit.
• get_granule(package_id, granule_id)

Auth

• Platform key: reuses PLATFORM_DATAGOV_KEY.
• BYO: ?_apiKey=<key> (api.data.gov key).

Data source

https://api.govinfo.gov — ?api_key= query param.

GAO appropriations law (fleet #1972)

Two tools that make GAO's secondary authority reachable by TOPIC rather than by package id.

gao_red_book(topic?, chapter?, include_scope_text?)

Resolves a fiscal-law question to the governing chapter of GAO's Principles of Federal Appropriations Law — the Red Book. Keyless: it reads www.gao.gov, not the GovInfo API, so it works with no data.gov key.

Why it does not use GovInfo. Measured 2026-09-14: the GAOREPORTS collection holds 16,569 packages but zero issued after 2010-01-01, and a full-text search for "Principles of Federal Appropriations Law" returns exactly one relevant package in its top 100 of 1,565 hits — the 2004 Update of the Third Edition. GovInfo cannot serve the current Red Book at all.

Edition handling is the point. GAO now publishes the Red Book chapter by chapter, so the current treatise is a mix:

"Our current Red Book consists of Chapters 1-3 (of the 4th Edition) and Chapters 5-15 (of the 3rd Edition)." — gao.gov, read live on every call

That sentence and the whole table of contents are parsed from GAO's page at call time rather than hardcoded, so Chapter 4 appears the day GAO posts it. Every row names its edition; the 2004 annual update is returned separately under superseded with the reason it is not the answer.

Freshness. source_last_modified is the Last-Modified of the chapter PDF itself. The TOC page's Last-Modified is a Drupal render timestamp that reads as today on every call — using it would tell a caller a 2019 chapter was revised this morning.

How matching works, and what it is worth. Chapter titles come off GAO's live page and are authoritative. On top of them sits RED_BOOK_TOPIC_ALIASES in src/index.ts — a list maintained in this pack, not something GAO publishes, which is why every matched row carries match_basis saying whether it matched title:, alias: or gao_scope:. The chapters are PDFs and GovInfo does not index them, so an alias cannot be verified against the text; entries are therefore restricted to terms that are the standard NAME of the rule the chapter's own title states (bona fide needs rule → Availability of Appropriations: Time), the US Code section that rule lives at, or a plain-English framing of the question the chapter's title asks. When nothing clears the score floor the tool returns match_confidence: "none" and the complete table of contents rather than a plausible wrong chapter.

gao_comptroller_decisions(query, date_from?, date_to?, include_reports?, page_size?)

GovInfo files GAO audit reports and Comptroller General decisions under one collection code — its own name says so, "Government Accountability Office Reports and Comptroller General Decisions" — so a relevance search interleaves them. Measured on "augmentation of appropriations": 100 hits examined, 15 of the top 25 were decisions; this tool returns the 26 decisions and drops the 74 reports, reporting both counts. The split is the package id (GAOREPORTS-B-…).

Coverage stops at 2009 — GAOREPORTS carries nothing newer. For recent bid protests use gao_protests_search; for the current rule use gao_red_book.

Quick Start

Add to your MCP client (Claude Desktop, Cursor, Windsurf, etc.):

{
  "mcpServers": {
    "govinfo": {
      "url": "https://gateway.pipeworx.io/govinfo/mcp"
    }
  }
}

What this endpoint actually serves

tools/list at https://gateway.pipeworx.io/govinfo/mcp returns the tools in the table above plus the shared Pipeworx meta-tools — ask_pipeworx, discover_tools, search_within, remember/recall and the rest of the gateway-wide set. So the tool count you see is larger than this table: a single-pack endpoint currently lists roughly 30 shared tools alongside the pack's own. The connection's initialize response states its exact scope, and is the authoritative answer for a given day.

This is deliberate, not multiplexing by accident. The meta-tools are what let a scoped connection answer a question this pack does not cover — via ask_pipeworx, which routes across the whole catalog — without you adding a second MCP server. There is currently no way to mount a pack endpoint without them; if the extra schemas cost you more context than the routing is worth, connect to the full gateway once rather than to several pack endpoints.

Or connect to the full Pipeworx gateway to get every pack's tools listed directly, instead of just this one's:

{
  "mcpServers": {
    "pipeworx": {
      "url": "https://gateway.pipeworx.io/mcp"
    }
  }
}

Both URLs reach the same gateway and the same 1679+ data sources. The only difference is which pack's tools are listed directly; ask_pipeworx reaches all of them from either one.

No MCP client? Call it over HTTP

curl -X POST https://gateway.pipeworx.io/v1/tools/govinfo_list_collections \
  -H 'Content-Type: application/json' \
  -d '{}'

No account needed for the first calls. Inspect any tool: GET https://gateway.pipeworx.io/v1/tools/govinfo_list_collections. Find one: POST https://gateway.pipeworx.io/v1/tools/search_packs with {"query":"..."}.

Standalone (no gateway account)

This package also runs as a local stdio MCP server — no Pipeworx account, no gateway round-trip:

{
  "mcpServers": {
    "govinfo": {
      "command": "npx",
      "args": ["-y", "@pipeworx/mcp-govinfo"]
    }
  }
}

Or run it directly to confirm it starts:

npx -y @pipeworx/mcp-govinfo

It speaks MCP over stdin/stdout and answers initialize/tools/list/tools/call for only this pack's tools — none of the shared meta-tools the gateway connection above adds. Same source, same tools, no ask_pipeworx routing.

Using with ask_pipeworx

Instead of calling tools directly, you can ask questions in plain English — this works on the pack endpoint above as well as on the full gateway:

ask_pipeworx({ question: "your question about Govinfo data" })

The gateway picks the right tool and fills the arguments automatically.

More

• Docs and guides
• pipeworx.io

License

MIT