Incremental OCR in pipeline builder

Im looking to run OCR in pipeline builder but am running into issues around scale. The input to my pipeline is a media set! Is there a good way to make sure that I am only running OCR on newly added files in some incremental fashion, or to just make sure that I am not re-running it on the same files.
Thank you!

Hey @apena! Right now this isn’t supported in Pipeline Builder for OCR