weather-apps

weather-apps

Demonstrates simple MCP (Model-Context-Provider) implementations for weather information retrieval using a client-server architecture and LLM integrations.

---

Projects Overview

This repository contains two example projects, each demonstrating a different approach to integrating LLMs and weather APIs using the MCP pattern:

• genai/: Uses Google Gemini LLM and OpenWeatherMap API.
• ollama/: Uses a local LLM via Ollama and the Open-Meteo API.

---

Project Details

1. genai/ — Gemini LLM + OpenWeatherMap

• Server: Flask REST API (src/weather_server.py) provides /api/v1/getWeather endpoint, fetching real-time weather from OpenWeatherMap.
• Client: Command-line tool (src/client.py) uses Google Gemini LLM to interpret user queries and call the weather server as a tool.
• LLM: Google Gemini (API key required).
• Weather API: OpenWeatherMap (API key required).
• Setup:
  • Install Python dependencies from requirements.txt.
  • Set up a .env file with Gemini and OpenWeatherMap API keys.
  • Start the server, then run the client.
• See genai/README.md for full setup and usage.

2. ollama/ — Local LLM (Ollama) + Open-Meteo

• Server: Python MCP server (src/weather_server.py) exposes a weather tool using the free Open-Meteo API (no API key required).
• Client: Command-line tool (src/client.py) uses a local LLM via Ollama (LangChain integration) to select and call the weather tool.
• LLM: Local LLM via Ollama (e.g., llama3, runs on your machine).
• Weather API: Open-Meteo (no API key required).
• Setup:
  • Install Python dependencies from requirements.txt.
  • Install and run Ollama, pull a model (e.g., ollama pull llama3).
  • Start the server, then run the client.
• See ollama/README.md for full setup and usage.

---

Comparison

Feature	genai	ollama
LLM	Google Gemini (cloud)	Local (Ollama, e.g. llama3)
Weather Source	OpenWeatherMap (API key)	Open-Meteo (free)
API Keys	Required	Not required (for weather)
Runs Locally	Partially (LLM is cloud)	Fully local (LLM & server)
Client-Server	Yes	Yes

---

Getting Started

• Choose a subproject based on your requirements (cloud vs. local, API key needs, etc.).
• Follow the setup instructions in the respective subproject's README.

For more details, see genai/README.md and ollama/README.md.