Mission and beliefs
Astral states that it builds high-performance developer tools for the Python ecosystem, with the aim of making that ecosystem more productive and enabling developers to ship software faster. The company describes a belief that a great tool can multiply the effectiveness of individual developers, teams and entire organisations. Its tools are built in the open, open source and permissively licensed, and the company says it strives to advance existing standards and integrate with the broader ecosystem. Astral names openness, action and craft as its values.
- Tagline: Next-gen Python Tooling
- Mission: make the Python ecosystem more productive
- Tools are open source and permissively licensed
- Stated values: openness, action, craft
Projects
Astral maintains three principal projects, each written in Rust and documented at docs.astral.sh. uv is a Python package manager that manages environments and dependencies, including for standalone scripts and for projects. Ruff is a Python linter and formatter distributed as a single CLI. ty is a Python type checker and language server, described as an alternative to mypy, Pyright and Pylance.
- uv — Python package manager written in Rust
- Ruff — Python linter and formatter written in Rust
- ty — Python type checker and language server written in Rust
Ruff linter and formatter
The Ruff linter runs via ruff check and is designed as a drop-in replacement for Flake8 and dozens of its plugins, as well as isort, pydocstyle, pyupgrade and autoflake. It supports over 900 lint rules, each re-implemented in Rust as a first-party feature, selected through the lint.select, lint.extend-select and lint.ignore settings using Flake8-style rule codes. The Ruff formatter runs via ruff format and is designed as a drop-in replacement for Black, intended to emit near-identical output on Black-formatted code; over extensive Black-formatted projects such as Django and Zulip, more than 99.9% of lines are formatted identically. Ruff is configured through pyproject.toml or ruff.toml, and a tutorial walks through adding the linter and formatter to a project created with uv.
- ruff check — linting, with --fix and --watch options
- ruff format — formatting, with --check and --diff modes
- Over 900 lint rules, enabled by default from the F, E, B, UP and RUF categories
- Compatible with Black when line-length settings match
Editor and CI integrations
The Ruff Language Server accepts settings that customise its behaviour and can read an existing pyproject.toml or ruff.toml file to configure the linter and formatter, with documented examples for VS Code, Neovim and Zed. Configuration may be given as a file path or as inline JSON, the latter introduced in Ruff 0.9.8, and the configurationPreference setting controls precedence between editor and project configuration. For continuous integration, Ruff can be installed from PyPI in a GitHub Actions workflow or run through the astral-sh/ruff-action action, which accepts version, args and src parameters. Ruff also documents GitLab CI/CD configuration producing a code quality report, use as a pre-commit hook via ruff-pre-commit, mdformat integration, and official Docker images.
- Ruff Language Server settings for VS Code, Neovim and Zed
- astral-sh/ruff-action for GitHub Actions
- GitLab CI/CD code quality report output
- ruff-pre-commit hook
- Docker images including alpine and debian variants
Running Python scripts with uv
uv executes Python scripts with uv run, managing script dependencies without manual environment handling and creating virtual environments on demand. Scripts without dependencies, or that use only the standard library, run directly, and arguments or input from stdin are supported. Inside a directory containing a pyproject.toml, uv installs the current project before running the script unless the --no-project flag is passed. The guide also covers declaring dependencies with inline metadata, shebangs, alternative package indexes, locking dependencies and selecting Python versions.
- uv run for standalone scripts
- Inline script dependency metadata
- --no-project flag to skip project installation
- Automatic virtual environment management
Team
Astral was founded by Charlie Marsh, who released Ruff before the company existed and describes the response to it as evidence of an opportunity to make the Python ecosystem more productive. The company describes itself as a small, distributed team of software engineers that is growing deliberately rather than rapidly. Its early team includes the authors of ripgrep, bat, hyperfine and maturin, early core contributors to Biome and Prefect, and multiple CPython core developers.
- Founder: Charlie Marsh
- Small, distributed engineering team
- Includes authors of ripgrep, bat, hyperfine and maturin
- Includes multiple CPython core developers
Community and contributing
Astral publishes its open-source repositories on GitHub and runs a Discord community where users can ask questions, share ideas and get support. The documentation site also links to news, updates and announcements from the team. Ruff has a dedicated contributing guide that labels issues as good first issue or help wanted, asks contributors to discuss new features before opening pull requests, and sets out an AI policy that all contributions must follow. Contributions to ty are covered by a separate ty contributing guide.
- Open-source repositories on GitHub
- Community Discord for questions and support
- good first issue and help wanted labels on Ruff issues
- AI policy applies to all contributions
Sources
Last verified 03 Nov 2026. This entry is compiled from the public web pages listed above. Nothing here is stated that those pages do not, and each of them was read on the date shown.