Skip to main content
Glama

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
US_PLACES_CONTACTNoAn email address or URL added to the User-Agent, so either service can reach you. Optional, and courteous.
US_PLACES_TIMEOUTNoHTTP timeout in seconds.60
US_PLACES_PLSS_URLNoBLM's national PLSS map service. Must be https; its host is one of the only two hosts the server will contact.https://gis.blm.gov/arcgis/rest/services/Cadastral/BLM_Natl_PLSS_CadNSDI/MapServer
US_PLACES_CACHE_DIRNoResponse cache directory.~/.cache/us-places-mcp
US_PLACES_OVERPASS_URLNoOpenHistoricalMap's Overpass endpoint. Must be https; its host is one of the only two hosts the server will contact.https://overpass-api.openhistoricalmap.org/api/interpreter

Instructions

Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.

This server publishes no instructions, or was last inspected before Glama recorded them.

Capabilities

Features and capabilities supported by this server

Protocol revision2025-11-25

CapabilityDetails
tools
{
  "listChanged": false
}
prompts
{
  "listChanged": false
}
resources
{
  "subscribe": false,
  "listChanged": false
}

Tools

Functions exposed to the LLM to take actions

NameDescription
county_atA

Which county held a point on a date, with the act that made it so.

Records were made by the county that held the place THEN: a 1795 deed for a farm now in Greene County, Pa., is in Washington County's books. Returns the holder(s), the full chain of counties for the point, and changes within a year of the date (boundary_change_within_a_year: check those by hand). A non-county area "attached to" a county was administered, and recorded, by that county. "contested" means two governments claimed the spot. The atlas ends in 2000 and holds no towns or townships. It says where the courthouse was, not where anyone lived.

county_historyA

Every version of one county: when it was created, from what, and each change.

Each version carries its dates, the event ("GREENE created from WASHINGTON."), and the statute. The atlas files counties under their modern state. Use county_at for a particular place: a county's history does not say which version held a given farm.

parse_legal_descriptionA

Read a Public Land Survey description into its parts. Makes no network call.

Handles GLO's padded 0840N/0390W, T84N R39W, E½NE / E2NE / "E 1/2 NE 1/4", words ("east half of the northeast quarter"), lots, and several tracts. Aliquots expand to quarter-quarters (E½NE -> NENE, SENE). A parse is a reading of a reading: raw keeps the text, and warnings says what was guessed (a bare section number, an inferred meridian). Pass a tract to plss_locate to place it on the map.

plss_locateA

Place a federal land description on the map, using BLM's PLSS data.

Returns each tract's centroid and bounding box (lat/lon), BLM's ids, and the precision reached (township, section, aliquot or lot). A tract is not a house: a section's centre is up to half a mile from any point in it, and BLM's modern survey data can differ from the original plat near correction lines and water. To find the county that held it, pass the centroid and the patent's date to county_at; a county today is not the county then.

plss_from_pointA

Name the Public Land Survey tract at a point: township, range, section, quarter-quarter.

Use it to turn a map pin (a cemetery, a farmstead) into the description a land patent or tract book would carry, then search the land records with it. Finds nothing in states that were never federal public domain.

public_land_stateA

Was this state federal public domain, and if not, who granted its land? No network call.

In a state that was never federal land (the 13 colonies, ME, VT, KY, TN, WV, TX, HI), a search for a federal patent finds nothing, and that nil says nothing about the family: first title came from the colony or state, and this says where those records are. Public-land states list their principal meridians, and the special cases (Ohio's surveys, Spanish and French claims).

find_land_entry_fileA

From a land patent to the case file behind it: which file, where, and the search to run.

The patent proves a conveyance on its signature date; the case file holds the application and, for a homestead, the proof of residence, witnesses and citizenship. Signature date is not purchase or settlement date: the entry came years earlier. For a military warrant the patentee may be an assignee, and the veteran is in the warrant file. Returns arguments for nara-catalog-mcp's search_records_advanced (this server cannot call it).

glo_linksA

Build a link to BLM's General Land Office Records site for the user to open. No network call.

GLO free text matches ANY word and ranks the results, so add a state and a category; there is no surname field. Links to the old site (before July 2026) now land on the home page: cite a patent by its accession number, document number, state and signature date, never by a link.

cache_statusA

Report this session's live calls and cache use. Makes no network call.

Prompts

Interactive templates invoked by user choice

NameDescription

No prompts

Resources

Contextual data attached and managed by the client

NameDescription

No resources

TDQS

A4/5.0

Scored across 9 tools

Disambiguation4/5

Tools are largely distinct: county_at (point-in-time jurisdiction) vs county_history (a county's versions) is explicitly differentiated, and plss_locate vs plss_from_point are inverse operations. There is mild overlap between parse_legal_description and plss_locate since both touch PLSS text, but the descriptions clarify that one parses and the other geolocates.

Naming Consistency4/5

All names use snake_case with a mostly predictable verb/noun or noun/noun structure (plss_locate, county_at, parse_legal_description). A few names are static noun phrases (county_history, cache_status, glo_links) rather than action-oriented, a minor deviation but still readable and consistent in casing.

Tool Count5/5

Nine tools is well-scoped for a land-research domain, with each tool covering a distinct capability (state context, PLSS geocoding both ways, county resolution, parsing, patent-to-file, link building, cache status). Nothing feels extraneous or missing at the count level.

Completeness4/5

The surface covers the PLSS-to-county research lifecycle well, including both directional geocoding, temporal county resolution, description parsing, GLO linking, and a hand-off to NARA for case files. Minor gaps like a direct patent/record lookup are delegated to companion servers by design, which is reasonable.

Maintenance

ActivityMaintained
ResponsivenessNo issues