Account and sync

After connecting an account, Resrch shows an initial inventory before the first sync cursor is committed. Choose:

• Initialize Metadata Sync to synchronize existing papers, publication metadata, authors, tags, collections, and notes while keeping PDF bytes local.
• Initialize with PDFs to synchronize the same metadata and upload eligible managed PDF files after their cloud paper records are created.

The inventory reports scanned, mapped, queued, and eligible papers together with the number and estimated size of managed-file candidates. Large libraries are processed in bounded, resumable batches. Re-running the action skips records that are already mapped or queued.

Incremental sync uses a durable local outbox and runs pull, push, then pull again. If a cursor becomes stale, Resrch requires bootstrap before further incremental pulls. Revision conflicts remain visible in the conflict center and are not silently overwritten.

Local files and cloud state remain separate boundaries. Metadata-only initialization does not upload PDF content.