About Address to Census Tract MCP
Address to Census Tract MCP lets AI assistants look up US census geographies from a street address or coordinates, using the US Census Bureau Geocoder. It also explains GEOIDs and supports current and specified census vintages without an API key.
Key features
- 01Look up geographies from US addresses
- 02Look up geographies from coordinates
- 03Return tract, block and district identifiers
- 04Explain GEOID components offline
- 05Choose a census vintage
Use cases
- 01Find a tract and GEOID for an address
- 02Identify districts for a coordinate
- 03Interpret a GEOID into FIPS parts
MCP support
Exposes address and coordinate lookup tools and an offline GEOID explanation tool to MCP clients. The hosted endpoint uses HTTP; the README also describes a planned local stdio option.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://address-to-census-tract-mcp.lipmichal.workers.dev/mcpclaude mcp add --transport http theluckystrike-address-to-census-tract https://address-to-census-tract-mcp.lipmichal.workers.dev/mcp{"mcpServers":{"theluckystrike-address-to-census-tract":{"type":"http","url":"https://address-to-census-tract-mcp.lipmichal.workers.dev/mcp"}}}Open in Cursor{"servers":{"theluckystrike-address-to-census-tract":{"type":"http","url":"https://address-to-census-tract-mcp.lipmichal.workers.dev/mcp"}}}Open in VS Code{"mcpServers":{"theluckystrike-address-to-census-tract":{"type":"http","url":"https://address-to-census-tract-mcp.lipmichal.workers.dev/mcp"}}}{"mcpServers":{"theluckystrike-address-to-census-tract":{"serverUrl":"https://address-to-census-tract-mcp.lipmichal.workers.dev/mcp"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
Tools
Details
From the project READMEaddress-to-census-tract-mcp
Address to census tract for AI assistants. Give a US street address or a latitude and longitude and get the exact census tract, block group, block, county FIPS and districts from the US Census Bureau Geocoder.
This is a Model Context Protocol (MCP) server. It calls the Census Geocoder at call time. It needs no API key.
Tools
| Tool | What it returns |
|---|---|
get_census_tract_for_address | The matched address, latitude and longitude, 11-digit tract GEOID and tract number, block group and block GEOIDs, state and county FIPS, incorporated place, county subdivision, congressional district, state senate and house districts, urban area and combined statistical area. |
get_census_tract_for_coordinates | The same geographies for a latitude and longitude. |
explain_census_geoid | A state, county, tract, block group or block GEOID split into its FIPS parts, with the tract number as Census shows it (012405 is 124.05). Works offline. |
Both lookup tools take an optional vintage. The default current uses today's geographies. census2020 uses the 2020 Census geographies, and acs2017 to acs2026 match an American Community Survey release.
Every result has structured JSON, a one-paragraph summary, as_of, source (the exact Census Geocoder request), a data.census.gov profile link for the tract and more (a page on mcpstamp.com).
What it knows that your model does not
- Which census tract a given address is in. Tract lines follow streets, rivers and rail lines, and the two sides of one street can be in different tracts. A model cannot work that out from an address.
- The 11, 12 and 15 digit GEOIDs that census data, CRA, LIHTC and grant eligibility lists are keyed on.
- The congressional and state legislative districts the Census Bureau currently publishes for that point.
Install
Hosted (Claude.ai custom connector, ChatGPT developer mode)
Paste this URL. No token, no sign-up. It works now.
https://address-to-census-tract-mcp.lipmichal.workers.dev/mcp
Claude Code
claude mcp add --transport http census-tract https://address-to-census-tract-mcp.lipmichal.workers.dev/mcp
Claude Desktop
Claude Desktop adds remote servers under Settings, Connectors, Add custom connector. Paste the URL above.
Cursor
Add to ~/.cursor/mcp.json:
{
"mcpServers": {
"census-tract": { "url": "https://address-to-census-tract-mcp.lipmichal.workers.dev/mcp" }
}
}
VS Code
Add to .vscode/mcp.json:
{
"servers": {
"census-tract": { "type": "http", "url": "https://address-to-census-tract-mcp.lipmichal.workers.dev/mcp" }
}
}
Local stdio (after npm release)
The npm package is not published yet. After it is, run it locally with Node 20 or newer:
claude mcp add census-tract -- npx -y address-to-census-tract-mcp
Example prompts
- "What census tract is 1 Dr Carlton B Goodlett Pl, San Francisco, CA 94102 in?"
- "Give me the 11-digit tract GEOID and block group for this address."
- "Which census tract and congressional district is 41.8827, -87.6237 in?"
- "What is the 2020 census tract for 1600 Pennsylvania Ave NW, Washington, DC?"
- "What does GEOID 060750124053007 mean?"
Data sources
Full list with fetch dates in DATA_SOURCES.md.
- Address matching and geographies from the US Census Bureau Geocoder, benchmark
Public_AR_Current. - State FIPS codes from the Census state reference file.
Limitations
- US addresses only.
- The Census Geocoder matches street addresses from Census address ranges. PO boxes, rural routes and very new streets may not match. Use the coordinates tool for those.
- When an address matches more than one record, the first match is used and the others are listed in
other_matches. - The geocoder can be slow. The tool waits up to 30 seconds and then returns a clear error instead of a guess.
More
More MCP servers that answer questions your assistant cannot answer alone are on mcpstamp.com.
Part of the Zovo MCP collection: https://mcp.zovo.one
Built by theluckystrike.
Development
npm install
npm test # builds, then runs the fixture tests and an offline stdio test
node scripts/stdio-transcript.mjs # live stdio round trip, writes .live/stdio-transcript.jsonl
License
MIT
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.