scikit-build

What it is and what it does

scikit-build acts as glue between Python's setuptools and CMake, letting you define native extension builds using CMake instead of setuptools' native build syntax. It reads your CMakeLists.txt and integrates the compiled artifacts into a Python package, handling platform differences and compiler selection automatically.

The package is designed for projects that already use CMake for their C/C++/Fortran/Cython code and want to package that code as a Python extension. It supports Python 3.8 through 3.14, modern Visual Studio versions including 2026, and handles cross-platform compilation details. The project is actively maintained with recent updates for compiler and setuptools compatibility.

Use it for:

• Build a library with C extensions defined in CMake and distribute it as a PyPI package.
• Compile Fortran code into a Python module while keeping your build logic in CMake.
• Integrate an existing C++ library into Python by wrapping it with Cython and using CMake.
• Support multiple platforms with a single CMake configuration for native extensions.
• Migrate a project from raw setuptools extension building to CMake without rewriting build logic.

Worth the install?

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

scikit-build bridges setuptools and CMake to compile CPython C/C++/Fortran/Cython extensions, handling integration between Python packaging and native code compilation.

Yes, if you are building Python C/C++/Fortran/Cython extensions and already use or prefer CMake. The package is actively maintained, has low install friction, carries no security vulnerabilities, and is permissively licensed. However, the description recommends evaluating scikit-build-core first if you do not require extensive build customization.

Install

scikit-build on PyPI

pip install scikit-build

uv add scikit-build

poetry add scikit-build

Installing scikit-build

Before you install

Low install friction with a pure-wheel distribution and only five runtime dependencies. Actively maintained with recent releases and ongoing support for current Python versions.

License in practice

Licensed under MIT (permissive), so you can use it freely in commercial and open-source projects without significant restrictions.

Quickstart

pip install scikit-build

Create setup.py with CMake configuration in CMakeLists.txt, then run:
python setup.py build

CMake 3.5 or later must be installed on your system; scikit-build is a build-time tool requiring a C/C++/Fortran compiler toolchain.

Verify before relying

• Whether scikit-build-core is recommended as a replacement for most new projects and migration guidance.
• Specific performance characteristics or build-time overhead compared to other build backends.
• Exact usage patterns and setup.py configuration required for typical extension projects.

Package facts

Evidence: scikit_build-0.19.1-py3-none-any.whl

Keywords: scikit-build

Tags

More Build Tools packages