Prototype · Preview: every server in the official registry, synced every 30 minutes. Reviews and accounts come in a later step.
PopNx MCPMCP servers for people who use AI · explained in plain words

US Places, Then

by iandersoversion 0.2.0i★ –iRuns on your computer
Registry nameio.github.ianderso/us-places-mcp

Install

Overview

Description from the maker (original English)

US places then: county on a date, PLSS land, GNIS names, old topo maps, post offices.

PopNx MCP’s three pillars

These three are never merged into one score: opinions, measurements and popularity are different things.

What you can do with it

We have not written a plain-language summary for this one yet. Below are the facts we read from the registry.

  • Connection: Runs on your computer
  • Transport: stdio
  • Category: Travel, maps & weather

Install

Pick the AI app you use, then copy the settings.

Before you install

  • Basic checks: 1 of 1 passed
  • Runs on your computer with your user’s permissions (files, network, apps).
  • You fill in yourself: US_PLACES_OVERPASS_URL, US_PLACES_PLSS_URL, US_PLACES_GNIS_URL, US_PLACES_GNIS_ARCHIVE_URL, US_PLACES_TNM_URL, US_PLACES_DATAVERSE_URL, US_PLACES_CACHE_DIR, US_PLACES_TIMEOUT, US_PLACES_CONTACT

You need

  • PyPIPython’s uv installed (for the uvx command)

Which app do you use?

Pick your app above. We remember it, so next time the button says “Install in …”.

Open Claude Desktop → Settings → Developer → Edit Config, and paste this into claude_desktop_config.json

{
  "mcpServers": {
    "us-places": {
      "command": "uvx",
      "args": [
        "us-places-mcp"
      ],
      "env": {
        "US_PLACES_OVERPASS_URL": "your-value",
        "US_PLACES_PLSS_URL": "your-value",
        "US_PLACES_GNIS_URL": "your-value",
        "US_PLACES_GNIS_ARCHIVE_URL": "your-value",
        "US_PLACES_TNM_URL": "your-value",
        "US_PLACES_DATAVERSE_URL": "your-value",
        "US_PLACES_CACHE_DIR": "your-value",
        "US_PLACES_TIMEOUT": "your-value",
        "US_PLACES_CONTACT": "your-value"
      }
    }
  }
}

Run this in your terminal

claude mcp add us-places --env US_PLACES_OVERPASS_URL=your-value --env US_PLACES_PLSS_URL=your-value --env US_PLACES_GNIS_URL=your-value --env US_PLACES_GNIS_ARCHIVE_URL=your-value --env US_PLACES_TNM_URL=your-value --env US_PLACES_DATAVERSE_URL=your-value --env US_PLACES_CACHE_DIR=your-value --env US_PLACES_TIMEOUT=your-value --env US_PLACES_CONTACT=your-value -- uvx us-places-mcp

Fill in the empty values before running it.

Open in Cursor

Or copy the settings yourself

Paste into .cursor/mcp.json in your project, or ~/.cursor/mcp.json

{
  "mcpServers": {
    "us-places": {
      "command": "uvx",
      "args": [
        "us-places-mcp"
      ],
      "env": {
        "US_PLACES_OVERPASS_URL": "your-value",
        "US_PLACES_PLSS_URL": "your-value",
        "US_PLACES_GNIS_URL": "your-value",
        "US_PLACES_GNIS_ARCHIVE_URL": "your-value",
        "US_PLACES_TNM_URL": "your-value",
        "US_PLACES_DATAVERSE_URL": "your-value",
        "US_PLACES_CACHE_DIR": "your-value",
        "US_PLACES_TIMEOUT": "your-value",
        "US_PLACES_CONTACT": "your-value"
      }
    }
  }
}

Your browser asks to open the app, then the app shows the settings and asks you to confirm. Key values are left empty for you to fill in.

Open in VS Code

Or copy the settings yourself

Paste into .vscode/mcp.json in your project

{
  "servers": {
    "us-places": {
      "type": "stdio",
      "command": "uvx",
      "args": [
        "us-places-mcp"
      ],
      "env": {
        "US_PLACES_OVERPASS_URL": "your-value",
        "US_PLACES_PLSS_URL": "your-value",
        "US_PLACES_GNIS_URL": "your-value",
        "US_PLACES_GNIS_ARCHIVE_URL": "your-value",
        "US_PLACES_TNM_URL": "your-value",
        "US_PLACES_DATAVERSE_URL": "your-value",
        "US_PLACES_CACHE_DIR": "your-value",
        "US_PLACES_TIMEOUT": "your-value",
        "US_PLACES_CONTACT": "your-value"
      }
    }
  }
}

Your browser asks to open the app, then the app shows the settings and asks you to confirm. Key values are left empty for you to fill in.

This one runs on your computer, and ChatGPT can only connect to servers on the internet. Pick another app.

Your choice stays on this device in a cookie; no account needed.

Values you fill in

  • US_PLACES_OVERPASS_URL env
    OpenHistoricalMap's Overpass endpoint.
  • US_PLACES_PLSS_URL env
    BLM's national PLSS (CadNSDI) map service.
  • US_PLACES_GNIS_URL env
    USGS's GNIS map service (The National Map).
  • US_PLACES_GNIS_ARCHIVE_URL env
    Where USGS keeps the August 2021 GNIS state files.
  • US_PLACES_TNM_URL env
    The National Map's TNM Access products API.
  • US_PLACES_DATAVERSE_URL env
    Harvard Dataverse, which holds the US post offices dataset.
  • US_PLACES_CACHE_DIR env
    Directory for the on-disk response cache and the downloaded datasets.
  • US_PLACES_TIMEOUT env
    HTTP timeout in seconds.
  • US_PLACES_CONTACT env
    An email address or URL appended to the User-Agent, so the services can reach whoever runs the server.

These settings are generated from the maker’s registry entry. If the README says otherwise, follow the README.

What it can do

Runs on your computer; the tool list appears in your AI app after you install it.

Basic checksi

We check information anyone can see; we have not read the code. Checks that do not apply are left out of the count. · Checked Nov 8, 2026

  • ✓ Pass

    Maker matches the repo

    The registry name was verified with the GitHub account ianderso, which owns the repo

  • … Not checked yet

    Still maintained

    We have not read this repo from GitHub yet

  • … Not checked yet

    Has a licence

    We have not read this repo from GitHub yet

  • – Not applicable

    Endpoint responds

    This one runs on your computer; there is no internet endpoint

  • – Not applicable

    Sign-in with OAuth

    Runs on your computer and uses local settings instead

  • … Not checked yet

    No hidden instructions in tool descriptions

    Runs on your computer; we can only read tool lists from internet endpoints

Compared with others in Travel, maps & weatheri

  • More stars than 0% of the other 1717 servers in this category
  • Published to the registry more recently than 49% of the other 1717 servers in this category
  • Basic checks equal to or better than 100% of the other 1717 servers in this category

Compared with every active server in this category in the registry. Stars count only when the publisher owns the repo.

Common questions

Is US Places, Then safe?

It passed 1 of 1 applicable basic checks. We have not read the code. Before adding a key or token, see what access the server asks for and use a key with only the permissions you need.

What keys or codes does it need?

The registry does not list a required key, though some servers ask you to sign in on first use.

Does it run on my computer or in the cloud?

It runs on your computer. Your AI app starts it when needed, and data does not pass through the maker’s servers (unless it calls an online service itself).

Is it still maintained?

The GitHub repo last changed –, and the latest version (0.2.0) reached the registry 32 days ago.

Which apps can use it?

Apps that support MCP, such as Claude Desktop, Claude Code, Cursor and VS Code. You also need Python (uv) on your computer.

What can I use instead?

In the same category: Cajon.pro, CarHauler247 Car Shipping Quotes, ThaiData UK

Alternatives

Servers in the same category from other makers, sorted by stars.

Reading a card:★ GitHub starsiThree pillars: reviews · checks · usageiUpdated = latest version published to the registryi

Cajon.pro

by cajon.pro

Drum schools map (Russia, Kazakhstan, Armenia, Belarus), song tempo and drum parts, cajons.

Runs in the cloudUpdated today
Install

ThaiData UK

by swifttiger888-commits

UK Thai restaurant data from FSA inspections: hygiene ratings, city stats, postcode search.

Runs in the cloudUpdated today
Install

README from the maker

Open the README

Copied from the maker’s repo as written. PopNx MCP has not reviewed or endorsed it. HTML, images and scripts are removed. Links lead off this site.

us-places-mcp

An MCP server for where things were, then. Records follow the jurisdiction that held a place on the date of the event, not today's: a 1795 deed for a farm now in Greene County, Pennsylvania, is in Washington County's books, because Greene was carved out the next year. This server answers the questions that decide which office to write to:

  • Which county held this spot on that date? From the Newberry Library's *Atlas of Historical County Boundaries*, which records every creation and boundary change of every US county from 1629 to 2000, dated to the day, with the act behind it. OpenHistoricalMap imported the atlas and serves it free (CC0); this server asks it.
  • Where is this land description on the ground? "E½NE Sec. 18, T84N R39W, 5th P.M." becomes a centroid, a bounding box and BLM's ids, from the Bureau of Land Management's national Public Land Survey data. A map pin becomes a land description the other way.
  • Where is the case file behind this land patent? The patent is the end of a process. The file at the National Archives holds the application and, for a homestead, years of residence, witnesses and citizenship papers. The server says which file and which series, and builds the search.
  • Where was Mount Hope Cemetery, or Hines Corners? USGS's Geographic Names Information System (GNIS) names and places every feature it knows, including vanished towns marked "(historical)". In 2021 it dropped cemeteries, churches, schools, post offices and other built features; the server also searches the archive USGS kept from August 2021, and says which answered.
  • Which post offices were there, and when? Richard Helbock's list of every US post office from 1639 to 2000, geocoded by Cameron Blevins: 166,140 offices with the years they opened and closed. A post office is often the only trace of a hamlet.
  • Which old maps show the spot? Every USGS topographic map since 1884 that covers a point, oldest first, with links to the scans.

Nothing here writes anywhere, and nothing here keeps a family tree. It sits well beside nara-catalog-mcp, which can run the archive searches this server builds, familysearch-mcp and snac-archives-mcp.

This is an independent project. It is not affiliated with, endorsed by, or supported by the Newberry Library, OpenHistoricalMap, the Bureau of Land Management, the US Geological Survey, Harvard Dataverse, the dataset's authors or the National Archives.

Tools

The server publishes twelve tools, all read-only and annotated so for the client. Five make no network call at all.

Jurisdiction at a date

ToolPurpose
county_atThe county (or counties) that held a point on a date: a day, a month or a year. Returns the holder, its dates, the event and statute that made it, the whole chain of counties for that spot, changes within a year (check those by hand), and whether two governments contested it.
county_historyEvery version of one county: created when, from what, and each later change, with statutes.

Federal land

ToolPurpose
parse_legal_descriptionRead a land description, however it is written (E½NE, E2NE, E 1/2 NE 1/4, "the east half of the northeast quarter", GLO's padded 0840N, lots, several tracts), into its parts, with quarter-quarters and warnings for anything guessed. Offline.
plss_locatePlace a description on the map: centroid, bounding box, acreage and BLM's ids, to the township, section, quarter-quarter or lot.
plss_from_pointName the survey tract at a point: township, range, section and quarter-quarter.
public_land_stateWas this state federal land? If not, who granted first title and where those records are; if so, its meridians and special cases. Offline.

Vanished places and old maps

ToolPurpose
find_place_nameA named place in GNIS: its point, class, state and county. Searches the live GNIS and, given a state, the archive of August 2021 that keeps the cemeteries, churches, schools, post offices and buildings GNIS dropped that year; says which answered.
post_officesWhich US post offices existed, 1639-2000: by name, by state and today's county, or around a point, and in a given year. Each with its years, whether it ran continuously, and its point if it was geocoded.
historical_topo_mapsEvery USGS topographic map covering a point, oldest first: date, scale and quadrangle, with GeoPDF, GeoTIFF and JPEG preview links, and a TopoView link for JPEG and KMZ.

From patent to case file

ToolPurpose
find_land_entry_fileFrom a patent's Authority: the kind of entry, what its file holds, which National Archives series has it, ready-made arguments for nara-catalog-mcp's search_records_advanced, and how to order the file if it is not online. Offline.
glo_linksA link to a search, or to one record, on BLM's General Land Office Records site, for you to open. Offline.
cache_statusThis session's live calls and cache hits, and the datasets downloaded so far. Offline.

Setup

You need Python 3.11 or later and uv. There is no key to request.

uvx us-places-mcp

or from a clone:

git clone https://github.com/ianderso/us-places-mcp

cd us-places-mcp uv sync uv run us-places-mcp # stdio server, usually launched by the client

Claude Desktop
{

"mcpServers": { "places": { "command": "uvx", "args": ["us-places-mcp"] } } }

If the server fails to start because uvx cannot be found, give the full path that which uvx prints as the command.

Claude Code
claude mcp add places -- uvx us-places-mcp

Configuration

Nothing is required. A .env file in the directory the server starts in supplies anything the environment does not; only that directory is read.

VariableMeaning
US_PLACES_OVERPASS_URLOpenHistoricalMap's Overpass endpoint. Default https://overpass-api.openhistoricalmap.org/api/interpreter.
US_PLACES_PLSS_URLBLM's national PLSS map service. Default https://gis.blm.gov/arcgis/rest/services/Cadastral/BLM_Natl_PLSS_CadNSDI/MapServer.
US_PLACES_GNIS_URLUSGS's GNIS map service. Default https://carto.nationalmap.gov/arcgis/rest/services/geonames/MapServer.
US_PLACES_GNIS_ARCHIVE_URLThe folder of GNIS state files from August 2021. Default https://prd-tnm.s3.amazonaws.com/StagedProducts/GeographicNames/Archive/MainDomestic.
US_PLACES_TNM_URLThe National Map's TNM Access products API. Default https://tnmaccess.nationalmap.gov/api/v1/products.
US_PLACES_DATAVERSE_URLHarvard Dataverse, which holds the post-office dataset. Default https://dataverse.harvard.edu.
US_PLACES_CACHE_DIRResponse cache directory; downloaded datasets go in its data folder. Default ~/.cache/us-places-mcp.
US_PLACES_TIMEOUTHTTP timeout in seconds. Default 60.
US_PLACES_CONTACTAn email address or URL added to the User-Agent, so the services can reach you. Optional, and courteous.

Every URL must be https. Their hosts, and Dataverse's file store (dvn-cloud-iqss.s3.amazonaws.com, where it redirects downloads), are the only ones the server will contact.

Being a good guest

OpenHistoricalMap's Overpass server runs on donated capacity and publishes no rate limit. The server sends one request at a time to each service, two seconds apart for OpenHistoricalMap, half a second for BLM and a second for USGS and Dataverse, joins identical calls in flight, and caches answers on disk: county answers for 90 days, GNIS and map answers for 30, survey answers until you pass refresh=true. A 429 or a 5xx is retried three times with back-off, then reported as rate_limited or upstream_error, which is never the same as "nothing here".

Downloaded data

Two sources are files rather than services, so the server downloads them once and reads them locally. Nothing is downloaded until a tool needs it, and the result that triggered a download says so: what, how big, how it was checked and where it was saved.

DataWhenSizeChecked against
US Post Offices (Blevins and Helbock, doi:10.7910/DVN/NUKCNA, CC0)The first post_offices call31 MB download, 21 MB on diskThe MD5 Dataverse publishes for the file
GNIS state file of 25 August 2021 (USGS, public domain)The first find_place_name search in that stateUp to 17 MB per state (Pennsylvania 9.8 MB); only the dropped classes are kept, about 6 MB for PennsylvaniaThe ETag USGS's bucket reports (an MD5, or the MD5 of the upload's parts)

Both are stored as SQLite files in data under the cache directory (~/.cache/us-places-mcp/data by default), never in the repository. A file that fails its check is deleted and the tool reports download_failed. cache_status lists what is there; delete the folder to start again.

How to read what comes back

  • The county then, not now. A record was made by the county that held the place on the day. A new county does not take its parent's earlier records, so a deed from before Greene County existed is in Washington County's books still.
  • Check dates near a change. boundary_change_within_a_year lists changes close to your date. Laws took effect on stated days, but offices took time to organise; records from those months can be in either county.
  • "Attached to". Before an area was organised as a county it was often attached to a neighbour for administration, and that neighbour kept its records. county_at reads this from the atlas's event text.
  • Contested ground. Pennsylvania and Virginia both governed the Waynesburg area in the 1770s, and similar disputes ran elsewhere. "status": "contested" means look in both governments' records.
  • The atlas files counties under today's state. Monongalia County was created by Virginia but is filed under West Virginia; the event text says which government acted.
  • The atlas ends in 2000 and holds counties only: not towns, townships or parishes as church units.
  • A located tract is not a house. A section's centre is up to half a mile from any point in it, and BLM's data is a modern compilation that can differ from the original plat near correction lines and water. Lotted sections (along a township's north and west edges, or by water) have lots where a regular section has quarter-quarters.
  • A parse is a reading of a reading. parse_legal_description keeps the original text and says what it guessed: a section number with no "Sec.", a meridian taken from the state.
  • A patent proves a conveyance, on its signature date. The entry was years earlier: homesteaders lived on the land five years before final proof, and mid-century backlogs put years between purchase and signature. The Homestead Act took effect in 1863, so an "1856 homestead" in a family story is a cash, credit, preemption or warrant entry.
  • The patentee may never have seen the land. Military bounty-land warrants were bought and sold; the veteran is in the warrant file, the patentee may be a speculator.
  • No federal patents in state-land states. The thirteen colonies, Maine, Vermont, Kentucky, Tennessee, West Virginia, Texas and Hawaii granted

Only the beginning is shown; read the rest in the repo

Ask PopNx