transform inputs: image/tiff
transform output: text/plain

It seems to add a lot of image/svg+xml entries too