Create a presigned direct-to-S3 upload slot.

Returns a short-lived presigned S3 PUT URL so a client can upload large binary artifacts (e.g. Revel Digital .webapp zip bundles) DIRECTLY to storage without the bytes streaming through the API. After uploading, call POST /media/uploads/{id}/finalize to ingest the object into the media library.

The account is always derived from the authenticated key/token — never from the body. The presigned URL is scoped to a single object key under the account's namespace, so it can never overwrite another account's objects.

Three-step flow:

# 1) Request a slot
curl -X POST "https://api.reveldigital.com/api/media/uploads" \
     -H "X-Reveldigital-Apikey: <API_KEY>" \
     -H "Content-Type: application/json" \
     -d '{ "name": "myapp.webapp", "group_id": "<GROUP_ID>", "tags": ["env=Production"], "content_type": "application/zip" }'
            
# 2) Upload the bytes straight to S3 (Content-Type MUST match the returned header)
curl -T myapp.webapp -H "Content-Type: application/zip" "<upload_url>"
            
# 3) Finalize / ingest
curl -X POST "https://api.reveldigital.com/api/media/uploads/<id>/finalize" \
     -H "X-Reveldigital-Apikey: <API_KEY>"

All media types supported by the multipart upload work here too (images, audio, video, presentations, PDF, and .webapp bundles): finalize downloads the staged object to the server and runs the SAME ingestion pipeline, so video is transcoded just as it is for a multipart upload.