jsonargparse

What it is and what it does

jsonargparse is a library for building command-line interfaces and making Python applications configurable without boilerplate. It reads your function or class signatures—their parameter names, type hints, and docstrings—and automatically generates a CLI that accepts arguments from the command line, YAML/JSON/TOML config files, or environment variables. It handles complex types (unions, optionals, nested dataclasses), dependency injection, and variable interpolation, all while keeping your core code free of CLI-specific decorators or inheritance.

The library is built on argparse but extends it significantly: it parses structured configs into nested hierarchies, supports multiple file formats, and can instantiate objects from parsed configurations. It's used as the foundation for pytorch-lightning's LightningCLI and is actively maintained with semantic versioning and full test coverage.

Use it for:

• Build a CLI tool that accepts both command-line flags and a YAML config file without writing separate parsing logic.
• Make a machine learning training script configurable via config files while keeping the training function clean and testable.
• Create a multi-step application where each step's parameters are defined as function arguments and automatically exposed to the CLI.
• Parse environment variables and config files into a structured dataclass without manual validation or type conversion.
• Support dependency injection in your CLI so users can swap implementations by changing config values.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

jsonargparse creates command-line interfaces and configuration systems from Python type hints and docstrings, parsing arguments from the command line, config files (JSON, YAML, TOML, Jsonnet), and environment variables.

Yes. jsonargparse is a solid choice for any Python project needing a CLI or configuration system. It has low install friction (one required dependency), active maintenance, no known vulnerabilities, and a permissive license. The design is non-intrusive—your code remains clean and testable. Use it if you want to avoid hand-writing argparse boilerplate or if you need config file support beyond what argparse provides.

Install

jsonargparse on PyPI

pip install jsonargparse

uv add jsonargparse

poetry add jsonargparse

Installing jsonargparse

Before you install

Low friction: pure Python wheel with only PyYAML as a required runtime dependency. Active maintenance with a release 23 days ago and consistent commit history. Supports Python 3.10 through 3.14.

License in practice

MIT license is permissive; you can use, modify, and distribute this package with minimal restrictions, making it suitable for both open-source and commercial projects.

Quickstart

pip install jsonargparse

from jsonargparse import auto_cli

def main(name: str, count: int = 1):
    for _ in range(count):
        print(f"Hello {name}")

if __name__ == "__main__":
    auto_cli(main)

Verify before relying

• Whether the 'substantial user base' claim is quantified beyond pytorch-lightning's LightningCLI adoption.
• Performance characteristics when parsing very large config files or deeply nested structures.
• Compatibility guarantees with optional dependencies (OmegaConf, shtab, argcomplete) across versions.

Package facts

Evidence: jsonargparse-4.50.0-py3-none-any.whl

Tags

More Application Frameworks packages