Weather Edge Mcp
About
Calibrated weather probability signals for Kalshi prediction markets. Dual-model: NWS forecast + GFS 31-member ensemble. Real-time METAR from settlement stations.
Details
- Author
- RJW34
- Downloads
- 143
- Categories
- Other, Finance, Media
Jump to
- Calibrates NWS daily high‑temperature forecasts by city
- Reads current Kalshi weather market prices
- Estimates per‑bucket probability, edge, and net expected value
- Exposes five MCP tools for AI agent consumption
- Ships an optional FastAPI web app with health and dashboard routes
- Supports multiple transports: stdio, SSE, and streamable HTTP
Setting up with Highlight
This MCP is not yet compatible with Highlight’s one-click setup. However, you can still use it with Highlight by following these steps:
- Download and install Highlight from highlightai.com/download
- Navigate to the plugins tab and select "Add Custom Plugin"
-
Configure the plugin with the settings below
Plugin Name
Weather Edge McpCommand (node, npx, python, etc.)Please refer to the README for specific instructions on how to obtain API keys or other required environment variables.
- Enable "Start Automatically" if you want the plugin to start when Highlight launches
From the repository
Install via pip install weather-edge-mcp. Configure it in Claude Desktop by adding the command python -m weather_edge_mcp to your MCP server list. For other clients, run weather-edge-mcp or python -m weather_edge_mcp with optional transport flags (--transport stdio, --transport sse --port 8050, or --transport streamable-http --port 8050).
Claude Desktop / Cursor
Paste into your MCP client config file to install this server.
{
"mcpServers": {
"weather edge mcp": {
"weather-edge": {
"command": "python",
"args": [
"-m",
"weather_edge_mcp"
]
}
}
}
}McpServers
{
"weather-edge": {
"command": "python",
"args": [
"-m",
"weather_edge_mcp"
]
}
}Weather Edge MCP Server
Weather Edge is an MCP server for calibrated Kalshi weather-market signals. It turns public forecast and market data into a compact tool surface for AI agents.
What it does
- calibrates NWS daily high-temperature forecasts by city
- reads current Kalshi weather market prices
- estimates per-bucket probability, edge, and net expected value
- exposes the results through MCP tools and an optional FastAPI surface
Install
pip install weather-edge-mcp
MCP usage
Claude Desktop
{
"mcpServers": {
"weather-edge": {
"command": "python",
"args": ["-m", "weather_edge_mcp"]
}
}
}
Other MCP clients
Use either of these commands:
weather-edge-mcp
python -m weather_edge_mcp
Transport options
weather-edge-mcp --transport stdio
weather-edge-mcp --transport sse --port 8050
weather-edge-mcp --transport streamable-http --port 8050
Tools
| Tool | Description |
|------|-------------|
| get_weather_signals(city) | Calibrated signals for one city's Kalshi weather markets |
| get_all_signals() | Full scan across all supported cities |
| get_forecast(city) | Bias-adjusted forecast context for one supported city |
| get_station_observation(city) | Latest METAR observation from the settlement station |
| list_cities() | Supported cities and calibration parameters |
Supported cities: nyc, chicago, denver, miami, la
Optional web API
Weather Edge also ships an optional FastAPI app:
python -m uvicorn weather_edge_mcp.web_app:app --host 0.0.0.0 --port 8080
Routes:
- /api/health
- /api/signals?city=nyc
- /api/all-signals
- /dashboard
- /subscribe
If the optional x402 stack is installed and configured, the paid routes can be gated there. MCP stdio mode stays clean and side-effect free.
Docker
The repo includes a Dockerfile for Glama/container builds.
docker build -t weather-edge-mcp .
docker run --rm weather-edge-mcp --help
Architecture
src/weather_edge_mcp/
core.py # forecasting, market fetches, calibration, formatting
mcp_server.py # MCP tools
web_app.py # optional FastAPI surface
cli.py # command-line entrypoint
Data sources
- National Weather Service forecast API
- Aviation Weather METAR API
- Kalshi public market API
Development
python -m unittest discover -s tests -v
python -m build
License
MIT
Sign in to leave a review
Use Google, GitHub, or an email account so ratings stay tied to real people.
No reviews posted yet.
