🐶 mcpup

Automatically generate Pydantic models for all functions in a Python package.

Features

• Automatic Function Discovery: Scans all modules in a package to find functions
• Pydantic Model Generation: Creates Pydantic models for function parameters using pydantic-function-models
• Validation: Generated models perform validation according to type hints
• Package Structure Preservation: Maintains the original package's module structure
• Optional uv Integration: Can install packages on-the-fly with uv

Installation

# Install with pip
pip install mcpup

# Or with uv
uv pip install mcpup

Requirements

• Python 3.10+
• uv (recommended)

Command Line Usage

Generate Pydantic models for all functions in a package:

mcpup package_name

Options:

--output, -o DIRECTORY       Directory to save generated models [default: ./mcpup_models]
--install, -i                Install the package using uv before generating models
--include-private            Include private functions (starting with underscore)
--module, -m TEXT            Specific modules to include (can be used multiple times)
--help                       Show help message and exit

Examples

Generate models for all functions in the polars package:

mcpup polars

Generate models only for specific modules:

mcpup polars --module dataframe --module series

Install the package first, then generate models:

mcpup some-package --install

Include private functions:

mcpup mypackage --include-private

Programmatic Usage

You can also use mcpup programmatically:

from mcpup.scanner import scan_package
from mcpup.generator import generate_models
from pathlib import Path

# Scan a package for functions
functions = scan_package("mypackage", include_private=False)

# Generate models
output_path = Path("./models")
generate_models(functions, output_path)

Using Generated Models

After generating models, you can use them to validate function arguments:

# Import the generated model
from mcpup_models.mypackage.mymodule import MyFunction

# Validate function arguments
valid_args = MyFunction.model.model_validate({
    "arg1": "value",
    "arg2": 123
})

# Call the function with validated arguments
from mypackage.mymodule import my_function
result = my_function(**valid_args.model_dump(exclude_unset=True))

MCP Integration

mcpup can be used to generate JSON schemas from Python packages, making it perfect for integration with Model Context Protocol (MCP) servers. MCP servers provide a standardized way for AI models to discover and use tools without custom integrations for each service.

Using mcpup with MCP Servers

Generate Pydantic models with mcpup, then access the JSON schemas to create MCP-compatible tools:

>>> from mcpup_models.requests import api
>>> from pprint import pprint
>>> api.Get.model
<class 'pydantic_function_models.validated_function.Get'>
>>> pprint(api.Get.model.model_json_schema())
{'properties': {'args': {'default': None,
                         'items': {},
                         'title': 'Args',
                         'type': 'array'},
                'kwargs': {'default': None,
                           'title': 'Kwargs',
                           'type': 'object'},
                'params': {'default': None, 'title': 'Params'},
                'url': {'title': 'Url'},
                'v__duplicate_kwargs': {'default': None,
                                        'items': {'type': 'string'},
                                        'title': 'V  Duplicate Kwargs',
                                        'type': 'array'}},
 'required': ['url'],
 'title': 'Get',
 'type': 'object'}

How This Powers MCP Servers

MCP servers use JSON schemas to:

1. Define Tool Capabilities: Each function in a package becomes a tool with a well-defined schema
2. Enable Natural AI Interaction: AI can understand the schema and use tools correctly
3. Support Mode Switching: Use with execution for actual API calls, or schema-only for documentation

You can turn any Python package into a composition of MCP-compatible tools, allowing AI systems to:

• Discover available functions
• Understand parameter requirements
• Validate inputs before execution
• Generate proper API calls

This approach makes Python packages accessible to AI systems in a standardized way, without requiring custom integration work for each package.

Contributing

Contributions welcome!

1. Issues & Discussions: Please open a GitHub issue or discussion for bugs, feature requests, or questions.
2. Pull Requests: PRs are welcome!
   • Install the dev extra with pip install -e ".[dev]"
   • Run tests with pytest
   • Include updates to docs or examples if relevant

License

This project is licensed under the MIT License.