Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@mcp-creatorHelp me build an MCP server that searches Spotify playlists."
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
mcp-creator
Create, build, and publish Python MCP servers to PyPI — conversationally.
Install mcp-creator, add it to your AI assistant, and it walks you through the entire process: naming your package, scaffolding a complete project, building, and publishing to PyPI.
Install
pip install mcp-creatorSetup
Add to Claude Code (~/.claude/settings.json):
{
"mcpServers": {
"mcp-creator": {
"command": "mcp-creator",
"args": []
}
}
}Or for Cursor (.cursor/mcp.json):
{
"mcpServers": {
"mcp-creator": {
"command": "mcp-creator",
"args": []
}
}
}Tools
Tool | What it does |
| Load your persistent profile — setup status, project history. Called first every session. |
| Save setup state, usernames, and project history across sessions |
| Detect what's installed (uv, git, gh, PyPI token) — only walks through missing steps |
| Check if a package name is available on PyPI |
| Create a complete MCP server project from a name + description + tool definitions |
| Add a new tool to an existing scaffolded project |
| Run |
| Run |
| Initialize git, create a GitHub repo, and push the code |
| Create LAUNCHGUIDE.md for marketplace submission |
How It Works
Tell your AI what you want to build: "I want an MCP server that checks the weather"
It checks the name: calls
check_pypi_nameto verify availability on PyPIIt scaffolds the project: calls
scaffold_serverwith your tool definitions — generates a complete, runnable projectYou fill in the logic: replace the TODO stubs in
services/with your real API callsBuild & publish:
build_package→publish_package→ live on PyPIPush to GitHub:
setup_githubcreates a repo and pushes your codeSubmit to marketplace:
generate_launchguidecreates the submission file with your repo URL
What Gets Generated
For a project named my-weather-mcp with a get_weather tool:
my-weather-mcp/
├── pyproject.toml ← hatchling build, mcp[cli] dep, CLI entry point
├── README.md ← install instructions + MCP config JSON
├── .gitignore
├── src/my_weather_mcp/
│ ├── __init__.py
│ ├── server.py ← FastMCP + @mcp.tool() for each tool
│ ├── transport.py
│ ├── tools/
│ │ ├── __init__.py
│ │ └── get_weather.py
│ └── services/
│ ├── __init__.py
│ └── get_weather_service.py ← TODO: your logic here
└── tests/
├── test_server.py
└── test_get_weather.pyThe generated server runs immediately — stub services return placeholder data so you can test before implementing real logic.
Requirements
Python 3.11+
uv (for building and publishing)
Development
git clone https://github.com/gmoneyn/mcp-creator.git
cd mcp-creator
uv venv .venv && source .venv/bin/activate
uv pip install -e ".[dev]"
pytest -v