What are Python metaclasses and when to use them

Short answer

A metaclass is a class of a class. It controls class creation by intercepting the class statement, allowing you to modify attributes, enforce naming conventions, or register subclasses automatically.

Details

By default, type is the metaclass for all Python classes. You can override it by passing metaclass= in the class definition or by inheriting from a class that already uses a custom metaclass. Common use cases include ORMs (Django models), singleton enforcement, and automatic registration of plugin classes. Metaclasses interact closely with Python classes and objects because they run before the class object is fully formed. If you write custom descriptors, metaclasses are a natural companion for injecting them into class definitions. For simpler needs, Python decorators applied to classes are often easier to maintain than a full metaclass.

Tips

• Prefer class decorators over metaclasses unless you need to control class construction order.
• Keep metaclass logic transparent; unexpected side effects confuse users of your library.
• Combine metaclasses with __init_subclass__ (Python 3.6+) for lighter-weight subclass hooks.