Open-Meteo MCP (Rust)

A lightweight, high-performance Model Context Protocol (MCP) server providing weather, snow conditions, air quality, and location data via the Open-Meteo API.

Rust port of open-meteo-mcp-java — featuring a 26.4 MB Docker image with <100ms cold start and 258 comprehensive tests.

Status: ✅ v2.0.0 - Production Ready | 72% test coverage | Feature-complete parity with Java v2.0.2

---

✨ Highlights

• ⚡ Sub-100ms Cold Start — Lightning-fast initialization
• 📦 26.4 MB Docker Image — Lightweight distribution
• 🎯 11 MCP Tools — Complete weather & climate toolkit
• 🧪 258 Comprehensive Tests — Production-grade quality assurance
• 🔄 STDIO + SSE Transports — Flexible deployment options
• 🦀 Pure Rust — Type-safe, memory-efficient, no GC pauses
• 📊 Connection Pooling — High-performance HTTP client with retry logic
• 🌍 Open-Meteo API — Free, reliable weather data source

---

Quick Start

Prerequisites

• Rust 1.75+ (install)
• macOS / Linux / Windows

Installation

# Clone repository
git clone https://github.com/schlpbch/open-meteo-mcp-rust.git
cd open-meteo-mcp-rust

# Build
cargo build --release

# Binary location
./target/release/open-meteo-mcp

Usage

Claude Desktop

1. Copy the binary to a location in your PATH:

   cp target/release/open-meteo-mcp ~/.local/bin/

2. Add to Claude Desktop config (~/.config/Claude/claude-desktop.json):

   {
     "mcpServers": {
       "open-meteo": {
         "command": "open-meteo-mcp",
         "args": ["--transport", "stdio"]
       }
     }
   }

3. Restart Claude Desktop

HTTP (SSE Transport)

# Terminal 1: Start server
cargo run -- --transport sse --port 8888

# Terminal 2: Test
curl http://localhost:8888/health

Features

MCP Tools (11 total) ✅

• get_weather — Weather forecast with temperature, precipitation, wind
• get_snow_conditions — Snow depth, snowfall, mountain weather
• get_air_quality — AQI, pollutants, UV index, pollen
• search_location — Geocoding - search locations by name
• search_location_swiss — Swiss-specific location search
• get_weather_alerts — Weather alerts based on thresholds
• get_comfort_index — Outdoor activity comfort score (0-100)
• get_astronomy — Sunrise, sunset, golden hour, moon phase
• compare_locations — Multi-location weather comparison
• get_historical_weather — Historical weather data (1940-present)
• get_marine_conditions — Wave/swell data for lakes and coasts

MCP Resources (4 total) ✅

• weather://codes — WMO weather code reference
• weather://parameters — Available API parameters
• weather://aqi-reference — AQI scales and health recommendations
• weather://swiss-locations — Swiss cities, mountains, passes

MCP Prompts (3 total) ✅

• ski-trip-weather — Ski trip planning with snow conditions
• plan-outdoor-activity — Weather-aware activity planning
• weather-aware-travel — Travel planning with weather integration

Configuration

Environment variables (see .env.example):

HOST=0.0.0.0           # HTTP host (default: 0.0.0.0)
PORT=8888              # HTTP port (default: 8888)
API_BASE_URL=...       # Open-Meteo API base (default: https://api.open-meteo.com)
TIMEOUT_SECS=30        # Request timeout (default: 30)
LOG_LEVEL=info         # Logging level (default: info)
TRANSPORT=stdio        # "stdio" or "sse" (default: stdio)

Development

Build & Test

# Build
cargo build --release

# Test (258 tests)
cargo test --all-features

# Format check
cargo fmt --check

# Lint
cargo clippy --all-targets

# Coverage report
cargo llvm-cov --html  # Report in target/llvm-cov/html/index.html

Test Suite (258 tests)

• Library Tests (78): Configuration, error handling, types
• Phase 4 Tool Handler Tests (91): Parameter validation, boundary testing
• Phase 5 Service Layer Tests (89): Error conversion, configuration management, service orchestration

Project Structure

src/
├── main.rs          # CLI entry point
├── lib.rs           # Library root
├── config.rs        # Configuration
├── error.rs         # Error types
├── service.rs       # Core service
├── tools/           # Tool implementations
├── client/          # HTTP clients (Phase 1)
├── resources/       # MCP resources (Phase 3)
├── prompts/         # MCP prompts (Phase 3)
└── transport/       # Transport implementations (Phase 4)

Architecture

See spec/ADR_COMPENDIUM.md for architecture decisions.

Key Decisions:

• Async: Tokio 1.x runtime
• HTTP: reqwest with connection pooling
• MCP SDK: Official rmcp 0.3+
• Transport: STDIO (Phase 0), SSE (Phase 4)
• Testing: Unit + Integration, 72% coverage target

Development Roadmap

Phase	Deliverable	Status
0	Scaffolding, ping tool, CI	✅ Complete
1	HTTP clients for 5 APIs	✅ Complete
2	11 tool implementations	✅ Complete
3	Resources & prompts	✅ Complete
4	STDIO + SSE transports	✅ Complete
5	Testing & coverage (72%)	✅ Complete (258 tests)
6	Docker & CD pipeline	✅ Complete

Status: All phases complete | Feature-complete parity with Java v2.0.2

Comparison with Java Version

Aspect	Java v2.0.2	Rust v2.0.0
Binary Size	50 MB + JVM	26.4 MB (Docker)
Cold Start	2-5s	<100ms
Memory (idle)	~150 MB	50-100 MB
Dependencies	100+ (Maven)	~16 direct (Cargo)
Transport	REST + MCP + Chat	STDIO + SSE
Chat Endpoint	✅ Included	❌ Client-side only
Test Coverage	72% (426 tests)	✅ 72% (258 tests)
Architecture Docs	Basic	✅ Enhanced (mermaid diagrams)

Documentation

• CLAUDE.md — AI development guide
• ARCHITECTURE.md — System design with mermaid diagrams
• spec/ADR_COMPENDIUM.md — Architecture decision records (12 ADRs)
• Open-Meteo API — Weather data source
• MCP Protocol — Model Context Protocol specification

License

Apache License 2.0 — See LICENSE file

Contributing

See CONTRIBUTING.md for guidelines.

Support

• Issues: GitHub Issues
• Discussions: GitHub Discussions

---

Maintained by: @schlpbch Status: ✅ v1.0.0 - Production Ready Last Updated: February 2026