snowsure-mcp-server
v3.7.0
Published
MCP stdio bridge to SnowSure hosted server — full tool surface (ask_snowdata, get_destination, insights). SnowSure.ai
Downloads
110
Maintainers
Readme
SnowSure MCP Server — ski resort snow for Cursor, Claude & AI agents
Model Context Protocol server for real-time powder rankings, 14-day multi-model snow forecasts, and 500+ ski resorts via SnowSure.
Use the hosted Streamable HTTP MCP — no install required. Paste
https://www.snowsure.ai/mcpinto Claude.ai, ChatGPT, Cursor, or any other client that supports remote MCP. 39 tools + 12 resources, including ask_snowdata (grounded Q&A) and 4 ChatGPT Apps SDK UI widgets.This npm package (v3.7+) is a thin stdio bridge: it proxies every request to the hosted server, so it exposes the same full tool surface. Use it only when your client cannot connect to HTTP MCP.
- Docs: snowsure.ai/developers/mcp · llms.txt · OpenAPI
- Machine-readable server cards:
/.well-known/mcp-server-card(SEP-2127) · full tool card
Features (hosted HTTP MCP at snowsure.ai/mcp)
- 39 AI tools including ask_snowdata (Answer Engine Q&A)
- 12 resources including snowsure://destinations and 4 Apps SDK UI widgets
- 500+ ski resorts worldwide
- 7 weather models for forecast comparison
- Hourly updates (see methodology)
- Reads need no auth; account tools (favorites, snow alerts, lodging booking) use OAuth 2.1
Available Tools
| Group | Tools |
|-------|-------|
| Q&A & search | ask_snowdata ★, search_resorts, get_destination |
| Conditions & forecasts | get_snow_report, get_resort, get_resort_info, get_resort_photos, find_best_powder, compare_forecasts, get_weather_forecast, find_resorts_by_criteria, get_snow_history, get_regional_summary, get_southern_hemisphere_report, get_season_openings, get_webcam_status, compare_resorts |
| Trip planning | plan_ski_trip, plan_ski_road_trip, find_powder_trips |
| Ski passes | get_pass, find_pass_resorts, compare_passes |
| Safety & access | get_avalanche, get_operating_risk, get_road_access, get_road_cameras, get_road_weather |
| Intelligence & ML | list_insight_categories, get_insights, get_ml_trends |
| Account (OAuth) | save_resort, remove_saved_resort, list_saved_resorts, subscribe_alerts, list_alerts, unsubscribe_alerts, book_lodging, get_my_snow_report |
Available Resources
| URI | Description |
|-----|-------------|
| snowsure://snow-report | Current global snow report (top 50) |
| snowsure://resorts | All 500+ resorts with conditions |
| snowsure://regions | Region/country breakdown |
| snowsure://destinations | Multi-mountain hubs (Niseko, Chamonix, Aspen, …) |
| snowsure://insights/catalog | Intelligence card index |
| snowsure://trends/catalog | ML/AI trend endpoint index |
| snowsure://faq/index | Seed FAQ index (all locales) |
| snowsure://api-docs | API documentation |
| ui://widget/powder-list.html | ChatGPT Apps SDK — ranked powder list card |
| ui://widget/snow-report-card.html | ChatGPT Apps SDK — single-resort snow report |
| ui://widget/resort-info-card.html | ChatGPT Apps SDK — resort guide card |
| ui://widget/resort-gallery-card.html | ChatGPT Apps SDK — photo gallery carousel |
Installation
Recommended: hosted Streamable HTTP MCP (no install)
Most modern MCP clients (Claude.ai, ChatGPT desktop, Cursor, MCP Inspector) support the Streamable HTTP transport. Add this URL as a remote / HTTP MCP server in your client of choice:
https://www.snowsure.ai/mcpThat's it — no install, no local Node process, always tracking the latest release. The endpoint is public and stateless; reads need no auth.
Alternative: install the npm package (stdio bridge)
For clients that only support stdio MCP servers, run the bridge locally — it proxies to the hosted server with the identical tool surface.
For Claude Desktop
Add to your claude_desktop_config.json:
{
"mcpServers": {
"snowsure": {
"command": "npx",
"args": ["-y", "snowsure-mcp-server"]
}
}
}For Cursor
Add to your MCP settings:
{
"snowsure": {
"command": "npx",
"args": ["-y", "snowsure-mcp-server"]
}
}Manual Installation
# Clone and build
git clone https://github.com/mikeslone/snowsure-web.git
cd snowsure-web/mcp-server
npm install
npm run build
# Run
node dist/index.jsExample Queries
Once connected, you can ask your AI assistant:
- "What are the best ski conditions right now?"
- "Tell me about Niseko's snow conditions"
- "Find me powder in Japan"
- "Compare forecasts for Matterhorn"
- "Plan a ski trip to the Alps in February"
- "Which resorts have the deepest snow base?"
- "What's the weather forecast for Jackson Hole this week?"
- "Is Epic or Ikon better value for a 10-day Utah trip?"
- "What's the avalanche danger and chain control situation near Palisades?"
Development
# Install dependencies
npm install
# Run in development mode
npm run dev
# Build for production
npm run build
# Watch mode
npm run watchData Sources
- Weather Models: ECMWF, GFS, GEM, JMA, ICON, Météo-France, Met Norway
- Conditions: Hourly weather pipeline (aligned with NWP model refresh cadence)
- Historical: 30 years of snowfall records
- SnowSure Score: Verified rating combining depth, recent snow, forecast, and historical reliability
License
MIT License - see LICENSE for details.
Support
- Website: snowsure.ai
- API Docs: snowsure.ai/developers
- Email: [email protected]
