Unlike merging with Osmosis, loading data into JOSM cannot easily be performed on a server. Using JOSM would require a user to manually initiate every single merge process.

It’s a good suggestion if the number of files is small, but that’s not how I understood the original question.