cli-to-py¶

Turn CLI binaries into Python APIs for applications and LLM agents.

cli-to-py reads a command's help output, builds a schema, and returns a Python object where subcommands are methods and CLI flags are keyword arguments.

import asyncio
from cli_to_py import convert

async def main():
    git = await convert("git")

    result = await git.status(short=True)
    print(result.stdout)

    branch = await git.branch(show_current=True).text()
    changed = await git("diff", name_only=True, _=["HEAD~1"]).lines()

    errors = git.validate("commit", massage="fix")
    print(errors[0].suggestion)  # "message"

asyncio.run(main())

Why It Exists¶

LLM agents often need to coordinate command-line tools repeatedly: loop over files, calculate intermediate values, branch on results, retry failures, and combine several tools into one workflow.

Sequential tool calls are a poor fit for that. Code is a better control plane. cli-to-py makes CLIs available as Python functions so an agent can write code that coordinates them.

This works well with code-mode interpreters such as Monty: the interpreter runs agent-written Python, while the host application decides which converted CLI functions are available.

What It Provides¶

Async-first API with a sync mirror.
Pythonic dispatch: api.status(short=True) becomes git status --short.
Function-shaped CLI access for agent code-mode workflows.
Manual pre-spawn validation for typo-prone flags.
Streaming, timeouts, cancellation, cwd/env overrides, and interactive stdio.
Optional schema cache for faster repeated conversion.
Standalone wrapper generation with .py and .pyi output.
Zero runtime dependencies.

Documentation¶

Status¶

Alpha. The core runtime paths are tested, but help parsing is necessarily best-effort because CLI help formats are not standardized.