# Government Context

> Structured federal policy records, events, and analysis for people and AI agents.

MCP Streamable HTTP endpoint: https://mcp.governmentcontext.com
Authentication: browser OAuth with a free or paid Government Context account.

## Documentation

Every page under /docs is also served as Markdown: add `.md` to its URL
(for example https://www.governmentcontext.com/docs/concepts.md).

- [Overview](https://www.governmentcontext.com/docs) ([Markdown](https://www.governmentcontext.com/docs.md))
- [Concepts](https://www.governmentcontext.com/docs/concepts) ([Markdown](https://www.governmentcontext.com/docs/concepts.md))
- [Authentication](https://www.governmentcontext.com/docs/authentication) ([Markdown](https://www.governmentcontext.com/docs/authentication.md))
- [Plans and data access](https://www.governmentcontext.com/docs/plans) ([Markdown](https://www.governmentcontext.com/docs/plans.md))
- [Claude Code](https://www.governmentcontext.com/docs/connect/claude-code) ([Markdown](https://www.governmentcontext.com/docs/connect/claude-code.md))
- [Claude.ai](https://www.governmentcontext.com/docs/connect/claude-ai) ([Markdown](https://www.governmentcontext.com/docs/connect/claude-ai.md))
- [ChatGPT](https://www.governmentcontext.com/docs/connect/chatgpt) ([Markdown](https://www.governmentcontext.com/docs/connect/chatgpt.md))
- [Tools](https://www.governmentcontext.com/docs/tools) ([Markdown](https://www.governmentcontext.com/docs/tools.md))

## Agent skill

- [Government Context usage skill](https://www.governmentcontext.com/docs/skill.md): the tools' mental model and core workflows as ready-to-use agent instructions (Agent Skills format).

## Guides

- [Five policy questions your AI agent can answer today](https://www.governmentcontext.com/guides/five-policy-questions-your-ai-agent-can-answer): Five real questions a government affairs team can put to Claude, ChatGPT, or any MCP-capable AI agent once it's connected to Government Context, with the actual answers.
- [How to install Government Context](https://www.governmentcontext.com/guides/how-to-install-government-context): Install Government Context in any MCP client. This hub covers the setup for each one, starting with connecting the free MCP server to ChatGPT in the desktop app and on the web.

## MCP tools

- [gc_search](https://www.governmentcontext.com/docs/tools/gc_search): Find federal government records by keyword: rulemakings (proposed rules, notices), dockets, agencies, policy topics, organizations, grant opportunities, agency press releases, bills, legislators, committees, and hearings. The front door of Government Context — returns natural-key ids every other gc_ tool accepts, plus a followable flag for gc_follow. Rulemaking results default to currently open comment periods; set filters.open_for_comment to false only when the user explicitly asks for closed comment periods. Query mode (query present) returns the best matches across the requested types with no pagination: refine the query rather than paging. Browse mode (no query, exactly one type) pages deterministically — e.g. types: ["topic"] lists every followable topic. Use keywords, not sentences. Available on every plan.
- [gc_list_activity](https://www.governmentcontext.com/docs/tools/gc_list_activity): Chronological government activity — what has happened recently, or over a time range. Optionally scoped through a View (an agency, topic, organization, rulemaking document, bill, legislator, or committee found via gc_search) and narrowed to activity groups: rulemaking (documents posted, comment deadlines, organization comments), press, grants, legislative (bill actions, cosponsors, committee changes), lobbying, awards, fec. Scoped to a single record (a bill, a rulemaking) it returns that record's complete history; through broader Views it returns the tracked feed — every response's coverage field says which. Use gc_list_comments for deep comment filtering and the getters for current state.
- [gc_get_rulemaking](https://www.governmentcontext.com/docs/tools/gc_get_rulemaking): Deep context for one federal rulemaking document (proposed rule, final rule, notice): current state, regulatory stage, comment-period status and deadline, docket context, policy topics, and the comment-analytics overview (stance split, commenter types, net support). Use the id from gc_search. include: related_documents, fr_action for more. Comments themselves: gc_list_comments; corpus analytics: gc_analyze_comments.
- [gc_get_comment](https://www.governmentcontext.com/docs/tools/gc_get_comment): One full public comment on a rulemaking: commenter identity (organization or person), commenter type, stance, analysis summary, full text (capped), and attachments as links. Campaign/issue membership included on paid plans. Use comment ids from gc_list_comments or gc_list_activity.
- [gc_list_comments](https://www.governmentcontext.com/docs/tools/gc_list_comments): Slice one rulemaking's public-comment corpus: by stance (support/oppose/other), commenter type (trade_association, union, business, advocacy, government, academic, individual), campaign membership (a campaign id or "organic"), tagged issue, date range, or attachments. Organizations list first, then most recent. The flagship research tool — 'which trade associations oppose this rule' is one call. Campaign/issue filters are paid analytics.
- [gc_analyze_comments](https://www.governmentcontext.com/docs/tools/gc_analyze_comments): Structured corpus analytics for one rulemaking — the verifiable aggregates behind 'summarize the comments': stance_by_type (who supports/opposes, by commenter type), timeline (weekly stance volumes), campaigns (form-letter campaigns vs organic split), issues (what commenters argue about), position_map (organizations × issues with evidence quotes). Request sections individually. You narrate; this supplies the numbers. Paid feature — free plans get the overview aggregates.
- [gc_get_grant](https://www.governmentcontext.com/docs/tools/gc_get_grant): One federal grant opportunity from Grants.gov: deadlines, award ceiling/floor, estimated funding, eligibility, funding instruments, and policy topics. Paid dataset — free plans get existence and title. Ids from gc_search (types: ["grant"]).
- [gc_get_press_release](https://www.governmentcontext.com/docs/tools/gc_get_press_release): One agency press release: full text (capped), summary, publication date, and policy topics. Id is the "AGENCY/slug" pair from gc_search or gc_list_activity.
- [gc_get_agency](https://www.governmentcontext.com/docs/tools/gc_get_agency): One federal agency's profile: description, rulemaking/comment volume stats, and a preview of rulemakings currently open for comment. Follow the agency to monitor everything it does.
- [gc_get_topic](https://www.governmentcontext.com/docs/tools/gc_get_topic): One curated policy topic (e.g. artificial-intelligence, water-quality): its place in the taxonomy (parent, child topics) and how to see activity classified under it. Topics are the cross-agency View — follow one to monitor a policy area everywhere it appears.
- [gc_get_organization](https://www.governmentcontext.com/docs/tools/gc_get_organization): One organization's government-affairs profile: comment activity stats and stance split, active policy topics, lobbying summary (registrants/clients), federal award totals, and political-giving rollups. Money sections gate to the previous quarterly reporting period on free plans. Ids are org slugs from gc_search.
- [gc_get_bill](https://www.governmentcontext.com/docs/tools/gc_get_bill): One congressional bill: stage, latest action, sponsor, cosponsor split by party, recent actions preview, related bills, policy topics, and the congress.gov text link. Id format: congress/type/number (e.g. 119/hr/3076). Complete action history: gc_list_activity with the bill as scope.
- [gc_get_legislator](https://www.governmentcontext.com/docs/tools/gc_get_legislator): One member of Congress: chamber, party, state/district, current committee assignments, terms served, and sponsored-bill count. Id is the bioguide id (e.g. S000033) from gc_search.
- [gc_get_amendment](https://www.governmentcontext.com/docs/tools/gc_get_amendment): One congressional amendment: purpose, description, author (with bioguide id), latest action, the bill it amends, policy topics, and a compact text preview when mirrored. Use gc_get_amendment_text for the complete plain text across cursor pages. Id format: congress/type/number (e.g. 119/samdt/2137) — from a bill's amendments section (gc_get_bill) or gc_list_activity amendment.introduced items. Amendments can't be followed — follow the bill or the author instead.
- [gc_get_hearing](https://www.governmentcontext.com/docs/tools/gc_get_hearing): One congressional hearing or committee meeting: schedule and status, committee, witnesses (with organization links), related bills, policy topics, and video link when available. Id is the Congress.gov event id from gc_search (types: ["hearing"]) or gc_list_activity. Hearings can't be followed — follow their committee or topics instead.
- [gc_get_committee](https://www.governmentcontext.com/docs/tools/gc_get_committee): One congressional committee: chamber, current roster with roles (chair, ranking member), and subcommittees. Id is the committee system code (e.g. hsju00) from gc_search. Roster changes over time: gc_list_activity with the committee as scope.
- [gc_get_bill_text](https://www.governmentcontext.com/docs/tools/gc_get_bill_text): The mirrored full text of a bill in Markdown, converted from the official Congress.gov XML (headings, tables, and quoted amendatory text preserved). Omit "sections" for the whole document, paging with next_cursor; pass section ids (e.g. ["sec-101"]) to read only those — the way to work with a long act. List the ids with gc_list_bill_sections; simple resolutions have none and always return in full. Id format: congress/type/number (e.g. 119/hr/3076).
- [gc_list_bill_sections](https://www.governmentcontext.com/docs/tools/gc_list_bill_sections): The addressable sections of a bill's text — id, heading, its place in the title/division hierarchy, and size. Published only when every section in the official XML matched its rendered heading, so an id always addresses exactly that section; documents without enumerated sections return an empty list. Feed the ids to gc_get_bill_text.
- [gc_get_amendment_text](https://www.governmentcontext.com/docs/tools/gc_get_amendment_text): The complete mirrored text of one congressional amendment, returned as exact plain-text pages. Congressional Record amendment pages do not carry reliable bill-like section ids, so use next_cursor to continue a long text; page boundaries prefer paragraphs or headings and fall back safely for an unbroken block. Id format: congress/type/number (e.g. 119/samdt/2137).
- [gc_list_bill_amendments](https://www.governmentcontext.com/docs/tools/gc_list_bill_amendments): Every amendment offered to a bill, newest first: purpose, author, latest action, and whether its text is mirrored. This is a metadata list; read a mirrored full text with gc_get_amendment_text. Congress publishes amendment text as Congressional Record pages, so it is plain text rather than Markdown.
- [gc_follow](https://www.governmentcontext.com/docs/tools/gc_follow): Subscribe the authenticated user to a View — a rulemaking document, policy topic, agency, organization, legislator, bill, or committee — to monitor its future activity by email digest. Cadence: daily (default), weekly, or instant (paid). Idempotent: following again updates the cadence. Use ids from gc_search (items with followable: true).
- [gc_unfollow](https://www.governmentcontext.com/docs/tools/gc_unfollow): Remove the authenticated user's follow on a View. Idempotent — the result reports whether a follow existed. Re-following later restores any customized alert triggers.
- [gc_list_follows](https://www.governmentcontext.com/docs/tools/gc_list_follows): The authenticated user's own active follows: each followed View (type, id, title), its alert cadence, and when it was created. This never lists other users or follower counts.
- [gc_submit_feedback](https://www.governmentcontext.com/docs/tools/gc_submit_feedback): Report missing data, wrong data, capability gaps, or bugs in Government Context. Include user_goal and details whenever your user is comfortable sharing them — submissions directly shape which data sources and tools we build next. Omit anything your user prefers to keep private. Use categories: missing_data (a dataset or record we should carry), data_error (something looks wrong), capability_gap (a question the tools can't answer), bug, other.