One or more translation templates (green rectangles) may be defined for each domain. A translation template specifies translation procedures to extract citation metadata from a template webpage, which may be used to extract metadata from other target webpages as well.

In addition, URL path patterns (pink rectangles) may be defined, creating translation subgroups within domains.

Given a target webpage (pink box):

1. First, Web2Cit checks what domain it belongs to. In the example, the target https://example.com/blog/page belongs to domain example.com.
2. Then, Web2Cit checks if the target matches one of the URL path patterns. In the example, the path /blog/page matches the /blog/* pattern. Otherwise, a catchall pattern (`/**/*`) is used.
3. Next, Web2Cit tries to translate the target using all templates matching the same pattern, in order, until an applicable template is found. Otherwise, a fallback template is used.
4. Finally, a citation is returned.