When uploading information into KM, the data is always packaged as a document.

A document can consist of a URL to a web page to download, one or more files, including images, or just a string to be stored.

Regardless of source and format, a client stores documents.

When searching and asking questions, you can specify the scope of the search, e.g. whether to consider an entire index, just a document, or maybe just one of the files in a document.

One important aspect to highlight is that documents have one unique ID, that you can either assign or is automatically assigned by the system. This ID allows to update the same files, without the risk of duplicating memories, and to delete the information at any moment.