Python

Python's Windows installer delivers CPython, the reference implementation of the Python language, together with the tools most projects expect: pip for packages, venv for isolated environments, IDLE for small scripts and experiments, and the py launcher for choosing between installed versions. The same release is available as source code and as installers for other operating systems; this entry covers the 64-bit Windows installer published on python.org.

What the standard install includes

The default install adds the interpreter, the full standard library, the Tcl/Tk toolkit that IDLE and tkinter need, pip, and documentation shortcuts. The py launcher is a small Windows-specific program that reads version markers and registry entries so that py -3.14 or py -3 selects a particular interpreter. Python also publishes an embeddable package for bundling into applications and, separately, a Microsoft Store edition; those are different distribution channels with their own update behaviour.

Setting up a working environment

The installer offers a per-user installation that needs no administrator rights and a system-wide installation that does. Adding Python to the PATH is optional and is the single decision most likely to cause confusion later, because a Store copy or an older interpreter may already answer to the name python. The reliable pattern is to leave the global PATH alone, create a virtual environment per project with python -m venv .venv, activate it, and install dependencies into it with pip. Upgrades then touch only the interpreter a project selected.

Limits worth knowing

Packages installed with pip are visible only to the environment that owns them, so a globally installed library is invisible inside a virtual environment, which is usually intended but surprising the first time. Windows ships no compiler by default, so packages published only as source need build tools before pip can compile them. Some packages lag behind a newly released Python version, which is one reason the newest interpreter is not always the right choice for a project with strict dependency pins. The launcher only knows about interpreters it can discover through installation or the registry.

Who should install something else

If a project pins an older Python, an environment manager that can install and switch interpreters may be more convenient than running several official installers side by side. Data-science users often prefer a distribution that bundles scientific packages, and anyone who only needs to run one packaged application should not install a language runtime at all. Choose the official installer when you want a clean, vendor-maintained CPython and are willing to manage environments yourself.