Requirements:
- other
MCP Weather Tool
About
What is Weather MCP Client?
A minimal MCP client that connects to the WeatherAPI.com service and a custom Ollama model. It runs as a Python script managed by uv and is designed for users who want to experiment with MCP tool use via a local language model.
How to use Weather MCP Client?
Install dependencies with uv, copy .example.env to .env and add your WeatherAPI key, then create a custom Ollama model named llama3.1-tool:8b using the provided modelfile based on llama3.1:8b. Start Ollama with ollama serve and run the client with uv run main.py.
Key features of Weather MCP Client
- Requires only uv and a WeatherAPI key
- Uses a custom Ollama model for tool‑aware responses
- Single‑command launch via uv run main.py
- Lightweight, no IDE or editor integration
Use cases of Weather MCP Client
- Testing MCP tool‑calling with a local LLM
- Fetching real‑time weather data through a conversational interface
- Experimenting with uv based Python MCP clients
FAQ from Weather MCP Client
What dependencies are needed?
You need uv for Python dependency management and a free API key from weatherapi.com.
How do I set up the custom Ollama model?
Create a model named llama3.1-tool:8b using the modelfile included in the repository, based on the llama3.1:8b base model.
Can I use a different LLM or provider?
The README does not mention support for other models or providers.
Is there any configuration beyond .env and the model?
No. After copying the environment file and creating the Ollama model, the client is ready to run.
Does this client work in an IDE or editor?
No. It is a command‑line script run via uv; no editor integration is described.
Details
- Author
- rairai77
- GitHub stars
- 1
- Category
- other
- Repository
- rairai77/weather-tool-with-mcp
Requirements:
- uv for dependencies - weatherapi.com keySetup:
- copy .example.env to .env and add the weather api key - create a custom ollama model called llama3.1-tool:8b using the modelfile and llama3.1:8b as a baseUsage:
ollama serve
uv run main.py