The Common Gateway Interface (CGI) is a standard protocol for interfacing external application software with an information server, commonly a web server. This allows the server to pass requests from a client web browser to the external application. The web server can then return the output from the application to the web browser.

The way CGI works from the Web server's point of view is that certain locations (e.g. http://www.example.com/wiki.cgi) are defined to be served by a CGI program. Whenever a request to a matching URL is received, the corresponding program is called, with any data that the client sent as input. Output from the program is collected by the Web server, augmented with appropriate headers, and sent back to the client.

Because this technology generally requires a fresh copy of the program to be executed for every CGI request, the workload could quickly overwhelm web servers, inspiring more efficient technologies such as mod_perl or ASP that allow script interpreters to be integrated directly into web servers as modules, thus avoiding the overhead of repeatedly loading and initializing language interpreters. However, this is only applicable for high-level languages that need interpreters. Such overloads can be avoided by utilizing languages like C. By using C or similar compiled languages it is possible to reach higher efficiency levels, because such programs terminate their execution cycle faster than interpreted languages with less operating system overhead.

The optimal configuration for any web application will obviously depend on application-specific details, amount of traffic, and complexity of the transaction; a software engineer analyzes these tradeoffs to determine the best implementation for a given task and budget. Web servers often have a cgi-bin directory at the base of the domain, to hold executable files. The overhead of spawning new processes to compile the server code can be easily handled if the code is only changed occasionally.

One example is FastCGI, while others include programming accelerators that take a web script when initially called and store a compiled version of the script in a system location so that further requests for the file are automatically directed to the compiled code instead of invoking the script interpreter every time the script is called. When scripts are changed, the temporary accelerator cache can be emptied to ensure that the new script is called instead of the old one.