Usda Soil MCP server

@pipeworx/usda-soil

USDA-NRCS Soil Data Access over SSURGO — the detailed US soil survey, mapped by field crews at 1:12,000-1:63,360 and republished annually. Soil at a point, the named soil series inside a map unit with their properties and horizons, and arbitrary SQL over the full survey schema.

Part of Pipeworx — an MCP gateway connecting AI agents to 1679+ live data sources.

Tools

acreage, farmland class, survey area and its publication date.

• usdasoilmapunitatpoint(lat, lon) — map unit key, symbol, official name,

unit with percentage, taxonomy, drainage class, slope range, hydrologic group, land capability class and hydric rating; optionally the depth horizons with sand/silt/clay, organic matter, pH, available water capacity and Ksat.

• usdasoilcomponents(mukey, include_horizons?) — the soil series in a map

• usdasoilquery(sql) — read-only SELECT over the SSURGO schema.

Auth

Keyless.

Not the same as soilgrids

soilgrids is ISRIC's 250 m machine-learning prediction for the whole globe. This is the surveyed US ground truth: named series, official taxonomy, interpretation ratings — and it covers only the United States and its territories. Use soilgrids outside the US, or where a modelled continuous surface is wanted instead of a mapped polygon.

Data sources

{"format":"JSON+COLUMNNAME","query":""}.

• — POST

Traps this pack absorbs:

the column-name row. runSql zips them into objects.

• It is SQL Server. TOP n, not LIMIT n.
• JSON+COLUMNNAME returns an array of arrays, not objects — Table[0] is

entirely XML-namespace boilerplate, so slicing the raw body hides the one sentence that names the problem. summarizeErrorBody from @pipeworx/shared unwraps the ; a bad column now reads USDA Soil Data Access: 400 Invalid query: Invalid column name 'nosuchcolumn'. (from the upstream's XML error document).

• Errors come back as OGC XML with a 400, and the first 300 characters are

error — usdasoil_query says so explicitly rather than presenting an empty result as a finding.

• **A valid SELECT that matches nothing returns an empty Table, not an

publication date) is on sacatalog, not legend. farmlndcl is on mapunit, not component. Land capability on component is nirrcapcl/nirrcapscl.

• Column locations are not where you would guess. saverest (survey area

with WKT point(lon lat) — longitude first.

• Point geometry goes through SDAGetMukeyfromintersectionwithWktWgs84

Quick Start

Add to your MCP client (Claude Desktop, Cursor, Windsurf, etc.):

{
  "mcpServers": {
    "usda-soil": {
      "url": "https://gateway.pipeworx.io/usda-soil/mcp"
    }
  }
}

What this endpoint actually serves

tools/list at https://gateway.pipeworx.io/usda-soil/mcp returns the tools in the table above plus the shared Pipeworx meta-tools — askpipeworx, discovertools, search_within, remember/recall and the rest of the gateway-wide set. So the tool count you see is larger than this table: a single-pack endpoint currently lists roughly 30 shared tools alongside the pack's own. The connection's initialize response states its exact scope, and is the authoritative answer for a given day.

This is deliberate, not multiplexing by accident. The meta-tools are what let a scoped connection answer a question this pack does not cover — via ask_pipeworx, which routes across the whole catalog — without you adding a second MCP server. There is currently no way to mount a pack endpoint without them; if the extra schemas cost you more context than the routing is worth, connect to the full gateway once rather than to several pack endpoints.

Or connect to the full Pipeworx gateway to get every pack's tools listed directly, instead of just this one's:

Shortened. The full README is on GitHub.