I see, thanks a lot for the explanation and feedback! Then it seems there might be some use cases where a module can be useful. We would also like to use OME-Zarr as an intermediate format, it would be interesting to see how much we can push on parallel processing using Nextflow processes and Zarr chuncks. I think that publishing pipeline outputs in OME-* formats makes a lot of sense, and I see your point on the cost of copying raw image data, thats a killer. In our case, we had not yet had the need to process datasets of 10s of TB, but we often need to curate datasets of diverse size (a very large on might be 2-3 TB) before storing the data in our OMERO server, often it means changing the format to OME-TIFF (with res. pyramid for large images), and additional curation of metadata into the OME-XML, so we could use some of these “OME-*” modules in a “pre-processing” step before archiving the data. For this we use custom Nextflow pipelines, and so I was wondering if other labs (or nf-core pipelines) might also need modules for this.