{get 'name'} asks the web server for another page of this application and puts what comes back in place of the tag. Unlike {page}, which builds the page inside the current request, this is a real second HTTP request: the page runs in a fresh process, with its own variables, and sees nothing of the page that asked for it.

The page is fetched with &padInclude, so it renders bare - without its _inits.pad/_exits.pad wrappers. Every variable assigned on the tag, {get 'badge', $colour = 'red'}, goes along on the query string, where the other page finds it as a request value. The answer is inserted as a value: its braces are text and are not parsed again.

The fetch goes to the server's own address, never to the host a request names. Under the strict check a page that does not exist is an error before any request is made. Because it costs a round trip, {get} is for the case where the isolation of a separate request is the point; {page} is faster and shares the request. Remote addresses are fetched with {curl}; a stored content block is read with {content:name}.