OpenPublica

Search US local-government meetings: AI summaries, transcripts, and adopted plan documents.

CommunityOnlineNo sign-inGlobalFreeRead-only

What it can do

    What data it sees

    Do you need an account

    No: the server works without sign-in

    Search US local-government meetings: AI summaries, transcripts, and adopted plan documents.

    How to connect

    How to connect OpenPublica to Claude

    The server address to paste into Claude settings:

    https://api.openpublica.com/mcp

    1. Open Claude (claude.ai or the desktop app).
    2. Go to Settings → Connectors.
    3. Click Add custom connector.
    4. Paste the server address copied below into Remote MCP server URL and click Add.
    5. In a chat, click + → Connectors and switch the new connector on.

    Custom connectors are available on Free, Pro, Max, Team and Enterprise plans (Free is limited to one). On Team and Enterprise an organization Owner adds the connector first under Organization settings → Connectors.

    Authorization

    After you click Add, a sign-in window for the service opens. Sign in with your own account and approve access. Claude never sees your password.

    How to connect OpenPublica to ChatGPT

    The server address to paste into ChatGPT settings:

    https://api.openpublica.com/mcp

    1. Open ChatGPT in a browser (chatgpt.com). A Plus, Pro, Business, Enterprise or Edu plan is required.
    2. Turn on developer mode once: Settings → Apps → Advanced settings → Developer mode.
    3. Open Settings → Connectors and click Create.
    4. Fill in the form: Name (anything), Description (one line about what the service does), MCP server URL (copy it below).
    5. Under Authentication choose OAuth if the service requires sign-in, otherwise None. Click Create.
    6. In a new chat open + → Apps/Connectors and enable the connector.

    OpenAI has renamed this section before (Connectors → Apps/Plugins). If the label differs, search settings for "developer mode". On Business/Enterprise workspaces an admin must allow custom connectors first.

    Authorization

    On first use ChatGPT opens the service's sign-in window. Sign in and approve access.

    How to connect OpenPublica to Cursor

    The server address to paste into Cursor settings:

    https://api.openpublica.com/mcp

    Fastest: click Open in Cursor below and confirm the prompt.

    Manually:

    1. In Cursor open Settings → Cursor Settings → MCP and click Add new global MCP server.
    2. Paste the JSON copied below into ~/.cursor/mcp.json (per project: .cursor/mcp.json in the repo root).
    3. Save the file. The server appears in the MCP list; authorize it there if asked.

    Authorization

    If the service needs sign-in, an authorize button appears next to the server in the MCP list.

    Server tool list (15)

    Raw names from tools/list. Only developers need these.

    search_meetings_semanticSearch OpenPublica meeting summaries by meaning, exact terms, or both. The best default tool for "what has been said/decided about X" questions. Results carry a `relevance_score` (semantic, 0-1) and/or `keyword_score` (BM25) plus `highlights`, so you can triage matches by confidence. If the response sets `degraded_to_vector`, the keyword index was unavailable and results are semantic-only — do NOT treat them as exact-term matches. Narrow with `category_tag`/`topic` (get valid values from `discover_meeting_vocabulary`), a date range, or one or more governments. Returns a summary preview; call `get_meeting` with a returned `meeting_id` for the full summary or transcript.
    list_recent_meetingsBrowse meetings newest-first, optionally filtered by government, type, topic, or date. Use this for "what are the most recent meetings" or to page through a government's history. For topical questions, prefer `search_meetings_semantic`.
    get_meetingGet a single meeting's full AI summary and topics. Set `include_transcript=true` to also retrieve the complete transcript text for direct quotation. Returns an error object if the meeting is not found.
    search_meeting_transcriptFind exact quotes/passages inside one meeting's transcript. Use this to pull direct quotations without loading a whole (often very long) transcript: it fetches the transcript server-side, scans it, and returns only the matching passages with approximate `start`/`end` timecodes when available. Exact matches rank first, with a fuzzy fallback. Prefer this over `get_meeting(include_transcript=true)` when you only need specific quotes.
    find_local_governmentsDiscover which governments OpenPublica tracks near a coordinate. Returns nearby governments with their `account_id`, which you can pass to the other tools to scope results to a specific city or board. Note OpenPublica only covers a limited set of US localities; an empty result means none are tracked near that point yet.
    find_governments_by_nameResolve a place *name* to the governments OpenPublica tracks there. Use this when you know a city/state name but not its coordinates — it returns matching governments with their `account_id` (prefix matches first), which you then pass to the search/list tools. Prefer this over `find_local_governments` when starting from a name rather than a latitude/longitude.
    find_governments_by_zipResolve a ZIP code to nearby tracked governments. Convenience wrapper: geocodes the ZIP, then finds the governments OpenPublica tracks nearest that point. Returns their `account_id`s plus the resolved location. An empty `governments` list means none are tracked near that ZIP yet.
    get_government_infoGet a tracked government's profile: name, location, and meeting categories.
    get_government_statsGet meeting-activity stats for a government over a recent window. Returns total meeting count, a per-day activity map, and a breakdown of meeting categories — useful for "how active has X been lately?" questions.
    discover_meeting_vocabularyDiscover the valid category and topic filter values before filtering a search. Meeting `category_tag`s are opaque UUIDs and topics have no fixed list, so you cannot guess them. Call this first to get the real values, then pass a returned `category_tag` (from `categories[].tags`) or `topic` string into `search_meetings_semantic` / `list_recent_meetings`. Scope with a government and/or date range; calls without explicit dates default to a recent window (returned as `default_window_days`).
    list_government_documentsList a government's adopted plan and budget documents. Each entry carries `covers` — every general-plan element that document answers for. A city that publishes one bound general plan satisfies all eight with the same PDF, so do not conclude an element is unpublished because no entry is named after it. `status` says what the entry offers: `converted` is searchable and readable; `superseded` is a retired edition, still stored but not searchable; and **`failed` means we hold this document and could not read it — it is not evidence the city published nothing.** Cite `source_url` in that case; it is always the city's own copy. An `availability` block appears ONLY when something is wrong with a document, and `availability.reason` says what. No `availability` key means there is nothing to report.
    get_documentOne document's metadata. Use read_document_pages to read it. If an `availability` key is present, read it before quoting figures: its `problems` are transcription-lint findings, most often a footnote marker that may have been merged into the number it annotates. No such key means the document transcribed cleanly.
    read_document_pagesRead a page range of a document as markdown. Pages are transcribed from rendered pixels, so tables and scanned text come through as text. A page that could not be transcribed appears as a marked hole rather than being silently dropped — if you see one, say so instead of reading past it. Text is assembled from page windows, so `covered` may be slightly wider than what you asked for; cite the page numbers in `covered`, not `requested`.
    search_documentsSearch adopted plan and budget documents, returning page-cited excerpts. Each hit's `text` is a short passage around the match, not the full page — call read_document_pages with the hit's `first_page` to read it in context. Every hit carries `first_page`/`last_page` and `source_url`. Quote those when citing: the page number is what makes the claim checkable against the city's own copy. This is word matching, not semantic search — try the terms a plan would actually use. Only currently-served documents are searchable, so an empty result is not evidence the government has no such policy; check list_government_documents for what is held and what could not be read. A hit with `document_complete: false` comes from a document with transcription findings. Verify it against `source_url` before quoting a number from it.
    get_coverage_overviewGet OpenPublica's overall coverage: total governments, states, and meetings indexed. DC is tracked but is not a state, so it is reported separately as `includes_dc` — read the pair as "N states and DC".