The main reason to use the ID approach is to work with embeddings on a file-level. This makes for targeted queries when combined with file metadata stored in a database, such as is done by LibreChat.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results