Importing Tracks
The Library uses the local filesystem as the source of truth for desktop imports. The import pipeline scans the flat bounces/, stems/, and projects/ inbox folders, groups high-confidence files, and organises those groups into tracks/.
What happens during import
Section titled “What happens during import”- The app scans the local inbox folders.
- High-confidence groups can organise and sync automatically.
- A file that needs an assignment decision stays in its matching flat inbox folder.
- The Import modal reports status, including organised, skipped, failed, and unassigned counts.
- Open Files to assign or correct an unassigned file. The retired Tracks-review destination picker is not part of the current workflow.
An exact existing-track match can use that track group’s stable ID. A possible fuzzy match is held for Files instead of creating a speculative track folder.
Duplicate safety
Section titled “Duplicate safety”The importer uses a SHA-256 content identity check before removing a leftover inbox item or treating a candidate as an exact duplicate. A matching filename or size alone is not enough.
If identity cannot be proven, hashing fails, or the content differs, the item is preserved. A different-content collision is preserved at a collision-safe destination rather than being overwritten or silently deleted.
AI grouping
Section titled “AI grouping”AI review is optional and limited to ambiguous local candidates. If it is unavailable, times out, or is disabled, the import continues through the same conservative assignment gate; it does not restore the deprecated Tracks review dialog.