Interfacy¶
Interfacy is a CLI framework that turns Python functions, classes, and class instances into command-line interfaces. It derives the CLI from signatures, type annotations, and docstrings instead of making you define it twice.
Features¶
Generate CLIs from functions, classes, class methods, and class instances.
Nested subcommands and manual command groups with aliases.
Type-driven parsing from annotations, with support for custom parsers.
Model expansion for dataclasses, Pydantic models, and plain classes.
--helptext generated from docstrings.Highly customizable help output with multiple layouts, color themes, and configurable ordering.
Stdin piping support with configurable routing to parameters.
Optional tab completion via
argcomplete.
Installation¶
From PyPI¶
pip install interfacy
uv add interfacy
From source¶
pip install git+https://github.com/zigai/interfacy.git
uv add "git+https://github.com/zigai/interfacy.git"
Quick Start¶
from interfacy import Interfacy
def greet(name: str, times: int = 1) -> None:
"""Print a greeting."""
print(" ".join([f"Hello, {name}!" for _ in range(times)]))
if __name__ == "__main__":
Interfacy().run(greet)
$ python app.py Ada
Hello, Ada!
$ python app.py Ada --times 2
Hello, Ada! Hello, Ada!
By default, required non-boolean parameters become positional arguments and optional parameters become flags.
Agent Skill¶
The official AI agent skill can be installed from this repo.
npx skills add https://github.com/zigai/interfacy/tree/master/interfacy/.agents/skills/interfacy
or
uvx library-skills
Class-Based Commands¶
Classes become command namespaces. __init__ parameters live at the command level and public methods become subcommands.
from interfacy import Interfacy
class Calculator:
def __init__(self, precision: int = 2) -> None:
self.precision = precision
def add(self, a: float, b: float) -> float:
return round(a + b, self.precision)
def mul(self, a: float, b: float) -> float:
return round(a * b, self.precision)
if __name__ == "__main__":
Interfacy(print_result=True).run(Calculator)
$ python app.py --precision 3 add 1.2345 2.3445
3.579
Structured Parameters¶
Dataclasses, Pydantic models, and plain classes with typed __init__ parameters can be expanded into nested flags and reconstructed before execution.
from dataclasses import dataclass
from interfacy import Interfacy
@dataclass
class Address:
city: str
postal_code: int
@dataclass
class User:
name: str
age: int
address: Address | None = None
def greet(user: User) -> str:
return f"Hello {user.name}, age {user.age}"
if __name__ == "__main__":
Interfacy(print_result=True).run(greet)
$ python app.py --user.name Ada --user.age 32
Hello Ada, age 32
Manual Groups¶
Use CommandGroup when your command tree is not naturally rooted in one callable:
from interfacy import CommandGroup, Interfacy
def clone(url: str) -> str:
return f"clone:{url}"
class Releases:
def cut(self, version: str) -> str:
return f"cut:{version}"
ops = CommandGroup("ops", description="Operational commands")
ops.add_command(clone)
ops.add_command(Releases)
if __name__ == "__main__":
Interfacy(print_result=True).run(ops)
Interfacy CLI Entrypoint¶
Interfacy also ships a CLI that can run an existing function, class, or class instance directly from a module or Python file:
$ interfacy app.py:greet Ada
$ interfacy app.py:greet --help
$ interfacy package.cli:Calculator add 1 2
The entrypoint supports configuration via TOML, loaded from ~/.config/interfacy/config.toml or INTERFACY_CONFIG.
usage: interfacy [--help] [--version] [--config-paths] [TARGET] ...
Interfacy is a framework for building CLIs from Python callables.
positional arguments:
[TARGET] Python file or module with a function/class/instance symbol
(e.g. main.py:main, pkg.cli:App, pkg.cli:service).
... Arguments passed through to the target command.
options:
--help Show this help message and exit
--version show version and exit.
--config-paths print config file search paths and exit.
Use 'interfacy TARGET --help' to display the help text for the target.