Wasmer FastMCP Demo Server MCP

This MCP server runs a FastMCP application on Wasmer Edge, exposing a small set of tools and dynamic resources you can access with any MCP-enabled client. It demonstrates how to deploy a Python-based MCP server to Wasmer Edge and how to run a local development instance.

How to use

You can connect to the Wasmer Edge deployment from any MCP-enabled client by using the /sse endpoint. The server provides three capabilities you can invoke from your client: a simple arithmetic tool that adds two numbers, a dynamic greeting resource that returns a personalized message, and a prompt template that formats greetings in different styles. Run the local server to experiment on your machine, or deploy to Wasmer Edge to expose the remote endpoint for production or teamwork.

How to install

Prerequisites: you need Python installed on your system. You also require a Python virtual environment tool and access to the internet to install packages.

Step 1: Create and activate a virtual environment.

python -m venv .venv
source .venv/bin/activate

Step 2: Install the MCP package inside the virtual environment.

pip install mcp

Step 3: Run the MCP server locally using the provided entry point. This starts the server with the Streamable HTTP transport.

python main.py

Additional sections

Deployment to Wasmer Edge exposes your MCP server at the following public endpoint. Use the /sse path when testing from an MCP-compatible client.

Wasmer Edge deployment overview: export an mcp instance from your Python script as mcp and deploy the application to Wasmer Edge. The public URL follows this pattern: https://<your-subdomain>.wasmer.app/sse. Clients connect to this URL to interact with your MCP server.

Local runtime command summary: to run the server locally, you install dependencies, create a virtual environment, and start the Python script that defines the MCP app. The final start command you use in local runs is python main.py.

Example MCP connection configuration

{
  "mcpServers": {
    "wasmer_demo": {
      "command": "python",
      "args": ["main.py"]
    }
  },
  "env": []
}

Available tools

add

A simple tool that sums two integers, exposed as add(a, b) and usable from MCP clients.

greeting

A dynamic resource that returns a personalized greeting for a given name, exposed as greeting://{name}.

greet_user

A prompt template that formats greetings based on the requested style, exposed as greet_user(name, style).