who_reps-mcp
Given a U.S. street address, returns the people who represent it (U.S. Senators, House member, Governor, and state legislators) using free public data, with no API keys required for federal officials.
README
who_reps-mcp
Address → your elected officials, as an MCP server. A free, open replacement for the Google Civic Information Representatives API, which was shut down on April 30, 2025 and left a generation of civic apps broken.
Give it a U.S. street address and it returns the people who represent it — U.S. Senators, U.S. House member, Governor, and state legislators — by stitching together free public data: address → districts (Census) → officials (open congressional + state-legislator data).
Why this exists
When Google killed representativeInfoByAddress, the official guidance was: use
our Divisions API for an OCD id, then go buy officeholder data from BallotReady,
Cicero, or Ballotpedia. This rebuilds the free path so an agent can just answer
"who represents this address?"
Quick start
The federal tier needs no API keys — the Census geocoder and the
unitedstates/congress-legislators dataset are both free and key-less.
uv sync
# As an MCP server (stdio):
uv run whoreps-mcp
# Or try the data path from a terminal (no keys needed):
uv run whoreps-mcp lookup "1600 Pennsylvania Ave NW, Washington, DC 20500"
uv run whoreps-mcp lookup "2300 N Lincoln Blvd, Oklahoma City, OK 73105"
# 4 officials for 2300 N LINCOLN BLVD, ...: U.S. Senator James Lankford (R);
# U.S. Senator Markwayne Mullin (R); U.S. Representative Stephanie I. Bice (R, 5);
# Governor Kevin Stitt (R). State legislators omitted — set OPENSTATES_API_KEY ...
To add the state-legislator tier, get a free
OpenStates key, cp .env.example .env,
set OPENSTATES_API_KEY, and the same lookups will include state senators and
representatives.
Use it as an MCP server
{
"mcpServers": {
"whoreps": {
"command": "uv",
"args": ["run", "whoreps-mcp"],
"cwd": "/path/to/whoreps-mcp"
}
}
}
Tools
| Tool | What it does |
|---|---|
lookup_officials(address) |
The headline tool. Address in, full slate out (federal + state where available), each official with party, contact, term, source, and as-of date, plus OCD divisions and coverage notes. |
list_districts(address) |
Just the districts/divisions for an address — congressional + state-legislative districts and OCD-IDs (the piece Google's Divisions API gave, free and self-hosted). |
lookup_by_district(state, chamber, district) |
Skip geocoding: senate/house (federal) or upper/lower (state). |
get_official_details(official_id) |
Enriched detail for one official — committees + full contact where available. |
Each tool returns structured JSON plus a concise text summary.
Graceful degradation
This is a civic tool, so correctness and honesty matter:
- Federal always works, with zero keys. Missing tiers never fail the whole
request — you get what's available plus
coverage_notesexplaining any gap. - DC and territories return the non-voting delegate (no senators), and note that DC has a Mayor rather than a Governor and no state legislature.
- Bad/PO-box/new-construction addresses that don't geocode return an empty result with a note, not an error.
- Freshness is visible. Every official carries its
sourceandas_ofdate. Governors are a vendored, dated snapshot (see docs/governors.md); everything else is fetched live and cached.
How it works
address ─▶ Census geocoder ─▶ {lat/lon, congressional district, state-leg districts, OCD divisions}
│
┌──────────────────────┼───────────────────────┐
▼ ▼ ▼
congress-legislators OpenStates people.geo governors.json
(U.S. Senate + House) (state legislators) (vendored)
└──────────────────────┼───────────────────────┘
▼
merge + normalize -> OfficialsResponse (SQLite-cached)
Sources live behind sources/ so the whole flow is tested against recorded
fixtures — CI is fully offline.
Development
uv run pytest # offline suite (recorded fixtures)
uv run ruff check . # lint
If uv run whoreps-mcp ever reports No module named 'whoreps_mcp' (a known
editable-install quirk on some setups, e.g. paths with spaces), run with
PYTHONPATH=src uv run whoreps-mcp ....
See SPEC.md and BUILD_PLAN.md for the design.
License
MIT.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.