WSGI

WSGI, or Web Server Gateway Interface, is a specification that defines how Python web servers communicate with Python web applications or frameworks. It was created to standardize the interface between servers and apps, enabling server-agnostic deployment and easier reuse of code across projects.

The WSGI specification originated in the early 2000s, with PEP 333 introducing the original interface and PEP 3333 later updating it for Python 3 compatibility. The core idea is a simple, language-agnostic contract that separates the web server from the application logic, while allowing middleware to sit in between.

At runtime, a WSGI server calls an application object with two arguments: environ and start_response. The environ is a dictionary containing request data (such as REQUEST_METHOD, PATH_INFO, QUERY_STRING) and a set of WSGI-specific variables (prefixed with wsgi.). The application must return an iterable that yields byte strings, representing the response body. Before generating the body, the application should call start_response with a status line (like "200 OK") and a list of HTTP headers. An optional exc_info parameter can be used to handle errors.

Middleware components can wrap a WSGI application to modify requests or responses without changing the application code. Common WSGI servers include Gunicorn, uWSGI, Waitress, and mod_wsgi; popular Python frameworks such as Django, Flask, and Pyramid provide WSGI-compliant applications.

WSGI has facilitated portability and ecosystem growth by decoupling web frameworks from servers. A related development in the ecosystem is ASGI, which addresses asynchronous workloads and provides an alternative for modern, async-capable web apps.