MCP 服务器

mcp-usa-climate-risk

Model Context Protocol server for real-time US climate risk and home insurance metrics.

README

mcp-usa-climate-risk

Model Context Protocol server for real-time US climate risk and home insurance metrics.

</div>

Overview

mcp-usa-climate-risk is an open-source Model Context Protocol server that exposes real-time US climate risk data to LLM agents. It provides a single, robust tool — get_zip_climate_risk — that queries the RiskBeforeBuy API to return parcel-level hazard scores, estimated homeowners insurance premiums, premium-to-income strain, and state insurance market pressure for any valid 5-digit US ZIP code.

Built for Claude Desktop, Cursor IDE, Windsurf, and any MCP-compatible AI environment.

Installation

Prerequisites

Node.js >= 18.0.0
npm or pnpm
A supported MCP client (Claude Desktop, Cursor, Windsurf)

Clone & Build

git clone https://github.com/opendata-collective/mcp-usa-climate-risk.git
cd mcp-usa-climate-risk
npm install
npm run build

Client Configuration

Claude Desktop

Add the following to your Claude Desktop MCP configuration file:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "usa-climate-risk": {
      "command": "node",
      "args": ["/absolute/path/to/mcp-usa-climate-risk/dist/index.js"]
    }
  }
}

After restarting Claude Desktop, the get_zip_climate_risk tool will appear in your available tools.

Cursor IDE

Add the following to your Cursor MCP configuration (.cursor/mcp.json in your project root, or the global settings):

{
  "mcpServers": {
    "usa-climate-risk": {
      "command": "node",
      "args": ["/absolute/path/to/mcp-usa-climate-risk/dist/index.js"]
    }
  }
}

Windsurf

Add the following to your Windsurf MCP configuration:

{
  "mcpServers": {
    "usa-climate-risk": {
      "command": "node",
      "args": ["/absolute/path/to/mcp-usa-climate-risk/dist/index.js"]
    }
  }
}

Development Mode

To run the server in development mode with hot-reloading via tsx:

npm run dev

Tool Reference

`get_zip_climate_risk`

Fetches comprehensive climate hazards, estimated home insurance premiums, premium-to-income strain, and market carrier pressure metrics for any valid 5-digit US ZIP Code.

Parameters:

Name	Type	Required	Description
`zip`	`string`	Yes	5-digit US ZIP code (regex: `^\d{5}$`)

Response (HTTP 200):

Returns a formatted Markdown summary:

## Climate Risk Report — ZIP 28801

| Metric | Value |
|--------|-------|
| **Overall Risk** | high |
| **Est. Annual Premium** | $2,287 – $4,130 / year |
| **Premium Strain** | high |
| **Market Pressure** | high |

_Source: Risk Before Buy — Federal data (FEMA, NOAA, USGS, USDA, EPA)._

Error Responses:

Status	Meaning	Message
404	ZIP not found	"ZIP code not found in our climate registry."
429	Rate limited	"Rate limit exceeded for the public tier (5 req/min)..."
Network error	Connection failure	Graceful error message, never crashes

Data Sources

All risk metrics are derived from verified federal datasets:

Agency	Dataset	Coverage
FEMA	National Risk Index (NRI)	Flood, hurricane, wildfire, earthquake
NOAA	Storm Events Database	Historical severe weather claims
USGS	Earthquake Hazards Program	Seismic peak ground acceleration
USDA	Forest Service Wildfire Risk	Wildfire probability to communities
EPA	Climate Exposure Overlays	Environmental hazard proximity

API Rate Limits

The public tier allows 5 requests per minute per IP. For bulk lookups and production integrations, consult the RiskBeforeBuy API Documentation.

Contributing

Contributions are welcome from data engineers, climate scientists, and MCP ecosystem developers. Areas of interest:

Additional MCP tools (e.g., parcel-level flood zone lookup, insurance market deep-dive)
Alternative transport implementations (Streamable HTTP, SSE)
Client SDK wrappers for Python, Go, Rust
Test coverage and data validation scripts

Please open an issue or submit a pull request.