MCP Server Creator

You are working with a meta MCP server that can generate and configure other MCP servers on the fly. This tool helps you create new MCP server configurations, build Python code for them, manage resources and tools, and export runnable server files for quick deployment.

How to use

You run the MCP Server Creator as a local tool and connect it to an MCP client. Use it to create a new server profile, add tools that expose API-like actions, attach resources you need (static data or templates), generate complete Python code for the server, and save that code to a file. You can manage multiple servers in memory, inspect details for each, and generate ready-to-run server artifacts.

How to install

Prerequisites: you need Python 3.8+ installed on your system.

Install the MCP Server Creator package from PyPI.

Choose a run method and start the tool.

Run methods and basic workflow

# Install the package
pip install mcp-server-creator

# Run as a standalone CLI using the recommended runtime
uvx mcp-server-creator

# Or run it as a Python module
python -m mcp_server_creator

Notes on usage with an MCP client

When you connect to an MCP client, you can select or create a server profile, then call tools to perform actions like creating a new server, adding tools, or exporting code. The generated code can be saved to a Python file and run independently, enabling quick deployment and testing.

Additional content

To extend functionality, you can add new tools with custom parameters, define resources (static or template-based), and customize code generation. The tool suite provides programmatic APIs for creating servers, listing them, and retrieving details, as well as saving generated code to disk.

Examples and practical tips

Create a new server profile, then add a tool to fetch data from an API, including parameter definitions and a return type. Generate the Python code for that server and save it to a file for deployment.

Security considerations

Only expose tools and resources that you trust. Use explicit parameters and return types to avoid leaking sensitive data. When exporting code, review any embedded implementation strings to ensure no secrets are hard-coded.

Troubleshooting tips

If the tool fails to start, verify Python 3.8+ is installed, and that you are using a compatible runtime (uvx). Check that the package is installed in the active environment and that the runtime command you chose is available in your PATH.

Notes

You can manage multiple MCP servers in memory, inspect details for each, and export complete runnable code for deployment. For complex setups, start with an example server to understand the workflow.

FAQ and support

If you need further guidance, consider building a minimal Weather Service example to learn the sequence: create a server, add a tool, add a resource, generate code, and save to a file. This iterative approach helps you validate each step before expanding to production configurations.

Available tools

create_server

Create a new MCP server configuration with a given name, description, and version.

list_servers

List all server configurations currently kept in memory.

get_server_details

Retrieve detailed information about a specific server by its identifier.

add_tool

Add a tool to an existing server, supporting sync or async tools, with defined parameters and a return type, and automatic import management.

add_resource

Add a resource to an existing server, including static data or dynamic templates with parameters and MIME types.

save_server

Export and save generated server code to a Python file.

generate_server_code

Generate complete runnable Python code for a specified server.

create_example_server

Create a complete example server to demonstrate capabilities (e.g., a Weather Service).