Storage overview

By source

Reading every source in parallel…

Where should new data live?

Index in Dataverse, body in Blob. Filter / sort / join fields go in Dataverse ($40/GB/mo, queryable). Heavy content — transcripts, raw payloads, embeddings, files — goes in Blob ($0.018/GB/mo, ~2,000× cheaper) with just a vo_blobUrl pointer left on the row.
🗃

Dataverse

~$40 / GB / mo
Pick when
  • You'll filter, sort, or join the data
  • It has relationships / foreign keys
  • Each row is small (a few KB)
📦

Azure Blob

~$0.018 / GB / mo
Pick when
  • It's a file (image, video, PDF)
  • It's content (transcript, payload)
  • Only fetched by name, never queried

Capacity caps

Override the default per-source GB ceilings.
Loading caps…