Mcpserver

You can run a lightweight MCP server locally to expose a set of tools and prompts that you can access from MCP clients. This setup uses FastMCP to expose Python-based tools and a demo server with resources and prompts, making it easy to experiment, automate tasks, and integrate with other systems.

How to use

Choose how you want to run the MCP server. You have two local entry points: a minimal tool server and a demo server with resources and prompts. Start the server you need, then connect your MCP client to it using the appropriate configuration. The tools and endpoints exposed are designed to be straightforward to call from your client, enabling you to run operations like greeting users, performing arithmetic, or listing files.

How to install

Prerequisites you need before starting are Ubuntu 22.04 or newer, Python 3.12 or newer, and Git. Optional tooling can speed up dependency management.

Step 1: Install system prerequisites

sudo apt update
sudo apt install -y python3.12 python3.12-venv git

Step 2: Create and activate a virtual environment

python3.12 -m venv venv
source venv/bin/activate

Step 3: Install dependencies

pip install --upgrade pip
pip install "mcp[cli]>=1.25.0"

Step 3b: Optional enhanced setup with uv

curl -LsSf https://astral.sh/uv/install.sh | sh
uv venv venv
source venv/bin/activate
uv pip install "mcp[cli]>=1.25.0"

Step 4: Run the minimal tool server

source venv/bin/activate
python server.py

Step 5: Run the example server with resources and prompts

source venv/bin/activate
python main.py

Additional content

The two local entry points expose different capabilities. The minimal tool server provides the tools greet_user, add_numbers, and list_files. The example server adds a tool, a named resource, and a prompt to demonstrate richer interactions. You can connect your MCP client to either server depending on whether you need a simple tool surface or a more feature-rich demo.

Configuration and notes

If your MCP client supports a settings file, you can point it to your running server. Create a configuration block that references the Python runtime and the script you want to execute. Use absolute paths for the runtime and script if your client requires them.

Example MCP client configuration you can adapt to your environment is shown below.

{
  "mcpServers": {
    "my_mcp_server": {
      "command": "/absolute/path/to/your/project/venv/bin/python",
      "args": [
        "/absolute/path/to/your/project/server.py"
      ]
    }
  }
}

Troubleshooting and tips

If you need to switch machines, recreate the virtual environment and reinstall the MCP CLI tools. Keep your venv out of version control by using a proper ignore setup. If you want to share configurations, export the mcpSettings.json with the correct absolute paths for your new environment.

Available tools

greet_user

Returns a greeting string when given a name, demonstrating a simple callable tool.

add_numbers

Adds two integers and returns the result, illustrating a basic numeric operation tool.

list_files

Lists files in a directory, showing how to access filesystem information through the MCP interface.