Documentation index

Runtime behaviour

Normal service startup resumes interrupted processing when configured with a model URL. An empty workspace, a completed batch, or an explicit stop request remains idle. Opening the Control page never starts a new batch.

New Batch selects ordinary files first and processes them in groups of up to

  1. After those groups, remaining batch capacity is filled by matching immediate child directories. Their contents are scanned recursively, but nested directories are not separate batch selections. Each directory item is identified, matched and imported before advancing.

TMDB selection, catalogue saving, media/SRT moves, duplicate handling and source cleanup are implemented. The worker runs independently of browser refresh. Unresolved items remain available for review; unmatched directories are left unchanged. A completed batch remains visible until replaced or cleared.

Stop Batch requests a cooperative stop after the current operation. Clear Current Batch waits for processing to stop and removes only workspace records. TMDB ID and Replace Local Media provide the supported manual corrections in the Action column; there is no action dropdown or Process Batch button.

The explicit --run-batch diagnostic mode runs identification without the normal automatic matching/import pipeline.

See the batch overview, directory patterns, import and cleanup, and persistent workspace.