Hello, I am currently looking to handle a process which looks like the process in the image. I have a bot that creates a SignedUrl to upload a document directly to S3 in a dedicated bucket and a DocumentReference. The signedUrl is stored in an extension of the documentReference. I want the documentReference attachment url to be the url to the bucket, but unfortunately as soon as the DocumentReference is saved the url is changed to a presignedUrl. Is there a way to bypass this ? For information we are not sending the document directly to the medplum bot because the document could be 2Go, so we prefer that the frontend handles directly the upload. Thank you.

HI @thomabig , thanks for your patience. I beleive what is going on here is that our external URL downloader is automatically fetching and storing the referenced data https://www.medplum.com/docs/fhir-datastore/binary-data#create-binary-via-external-url