
MCP Server demo
This is a demo of an MCP server using Python, intended to be tested with vscode Copilot in Agent mode
What is MCP Server Demo?
MCP Server Demo is a demonstration of an MCP server built using Python, designed to be tested with Visual Studio Code (VSCode) in Agent mode.
How to use MCP Server Demo?
To use the MCP Server Demo, start a new Devbox shell, open the environment in the devcontainer, run the server, and set GitHub Copilot to Agent mode to interact with the server.
Key features of MCP Server Demo?
- Integration with GitHub Copilot in Agent mode
- Ability to process commands related to CSV and Parquet files
- Easy setup and testing within VSCode
Use cases of MCP Server Demo?
- Summarizing CSV files using GitHub Copilot.
- Querying the number of rows in Parquet files.
- Demonstrating the capabilities of an MCP server in a development environment.
FAQ from MCP Server Demo?
- What is required to run the MCP Server Demo?
You need VSCode version 1.99.0 or higher and a Devbox environment set up with the necessary packages.
- Can I use this demo for production?
This demo is intended for testing and development purposes only, not for production use.
- How do I set Copilot to Agent mode?
In the dropdown option next to the prompt textbox in VSCode, select "Agent" to enable Agent mode.
MCP Server demo
This is a demo of an MCP server and it's intended to be tested in vscode.
Based on Building a Basic MCP Server with Python by Alex Merced
Development
Start a new Devbox shell that has your packages and tools installed:
devbox shell
Testing it in vscode
Requirements
- VSCode Version =>1.99.0
Instructions
- Open the environment in the devcontainer (vscode command palette >
Dev Containers: Reopen in Container
) - Run
task mix-server:run
- Open the command palette and select
MCP: List Servers
- Choose
mix_server
- Start it
- Set Copilot to
Agent mode
by choosing "Agent" in the dropdown option next to the prompt textbox. - Ask Github copilot something like "Summarize the CSV file named sample.csv." or "How many rows are in sample.parquet?" Copilot will detect the appropriate tool, call the MCP server, and respond with the results—powered by the Python code.