MCP servers can provide the functionalities like:
- Research: File-like data that can be ready by client (like api responses or file contents)
- Tools: Functions that can be called by the LLM (with user approval)
- Promots: Pre-written templates that help users accomplish specific tasks
- Python 3.10 or higher
- uv as package manager
- Python MCP SDK 1.2.0 or higher
-
Install uv if not available {curl -Lsf https://astral.sh/uv/}
-
Initialise the project with uv
-
Add the dependencies {uv add "mcp[cli]" httpx}
-
Create weather.py file which has content for the server
-
Run the Script using npx @modelcontextprotocol/inspector -- "uv" "run" "weather.py" or uv run weather.py