# landbenchmark-mcp [Health: Active]

**Category:** 🚀 Aerospace & Astrodynamics  
**Repository:** https://github.com/winds3753/landbenchmark-mcp  
**GitHub Stars:** 0  
**npm Downloads (last month):** 90  
**Views:** 1  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/landbenchmark-mcp

## Description
Satellite land due-diligence: a green/caution/walk-away verdict with cited signals for any parcel.

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "landbenchmark-mcp": {
    "command": "npx",
    "args": ["-y","landbenchmark-mcp"],
    "env": {
      "TERRAIN_API_KEY": "",
      "TERRAIN_BASE_URL": ""
    }
  }
}
```

**Requires environment variables:** `TERRAIN_API_KEY`, `TERRAIN_BASE_URL` — the values above are empty placeholders; fill in real credentials before running (see the repository for what each one is for).

## Documentation

## What landbenchmark-mcp MCP server does

The landbenchmark-mcp MCP server connects an MCP-capable AI client to LandBenchmark, a land-analysis service that uses satellite data for parcel due diligence. It is intended for questions such as whether a parcel appears suitable for further investigation or purchase review. The result is a plain-text assessment containing a verdict, cited signals, and a URL for the full report.

The service evaluates several land-related factors: flooding, slope and buildability, soil, hazards, and access. Its output uses one of three classifications: green, caution, or walk-away. These classifications summarize the available signals; they are not a survey, a formal flood determination, or a replacement for inspection or professional advice.

## How it works

An AI agent calls the `analyze_parcel` tool with either a parcel centre as WGS84 latitude and longitude or a GeoJSON geometry. The geometry can be a point or polygon. An optional label can identify the parcel in the request or resulting report.

The tool supports two analysis modes. `lite` is the default and typically takes about one minute. `full` performs a deeper, multi-year satellite analysis and generally takes three to four minutes. The returned summary includes flagged signals and a link to the complete LandBenchmark report.

## Setup and configuration

To run the landbenchmark-mcp MCP server, obtain an API key from the API keys section of the LandBenchmark account site. Configure the server as an MCP process launched with `npx -y landbenchmark-mcp`, then provide both required environment variables:

- `TERRAIN_API_KEY`: the LandBenchmark API key.
- `TERRAIN_BASE_URL`: the HTTPS base URL for the service, normally `https://www.landbenchmark.com`.

The base URL is deliberately required rather than assigned a default. Each request sends the API key to that host in an `Authorization` header. The server rejects plain HTTP destinations except for `localhost`, which prevents sending the key to an unsecured remote endpoint.

The repository can also be built with `npm install` followed by `npm run build`; tests are run with `npm test`, including a check for the API-key safety guard.

## Tools and capabilities

The landbenchmark-mcp MCP server exposes one tool, `analyze_parcel`. Its supported inputs are:

- `lat` and `lon`: numeric WGS84 coordinates for the parcel centre.
- `geometry`: a GeoJSON point or polygon, used instead of coordinates.
- `label`: an optional parcel name.
- `mode`: `lite` or `full`, with `lite` as the default.

Coordinates or geometry are required as the parcel location. The tool returns the verdict summary as plain text, including the signals that influenced the result and a link to the full report.

## Limitations and notes

Results are based on observed satellite data and should be treated as informational. They do not constitute a land survey, flood determination, on-site inspection, or professional advice. The full mode is slower than lite mode, so it is better suited to cases where a deeper multi-year report is worth the additional wait.

The server works with MCP-capable clients, including Claude Desktop and Claude Code. Because it is configured as a standard command-based MCP process, it can also be used with other clients that support stdio MCP servers. The API key and base URL must be configured before startup; the server intentionally refuses to guess the destination host.

_Full upstream README: https://allmcps.com/mcp/landbenchmark-mcp/readme_