It is impossible to 1:1 replicate the input as an output because the images are not stored. It isn't a database. It's basically aggregating summaries/abstractio

It is impossible to 1:1 replicate the input as an output because the images are not stored. It isn't a database. It's basically aggregating summaries/abstractions/generalizations of a bunch of tags. In other words, it is transformative by default.