The content crawler's click-through implementation must return content in a readable format and set the content type and file name using the appropriate headers.
The following example uses a file, but the crawled resource could be any type of content. If the content is not in a file, the click-through servlet should create a representation with as little extraneous information as possible in a temporary file (for example, for a database, you would retrieve the record and transform it to HTML). See Creating Temporary Files for Indexing. You can also use the Oracle WebCenter Interaction Development Kit (IDK) DocFetch mechanism to handle indexing and click-through; see Implementing Content Crawler DocFetch.