Product methodology
How TrackRescue detects and repairs unavailable playlist entries.
TrackRescue separates observation, identification, candidate evaluation, playlist writing, and undo into distinct stages. This prevents a count mismatch or similar title from being treated as sufficient proof for an automatic playlist change.
Evidence model
How evidence is evaluated
Complete playlist pagination
A scan follows every next-page token so large playlists are not assessed from only the first group of entries.
Stable identifiers
Playlist item IDs, video IDs, playlist IDs, and positions are stored separately from human-readable titles.
Availability classification
Deleted, private, restricted, identifiable unavailable, and anonymous hidden gaps are recorded as different observations.
Snapshot comparison
The current row is compared with previously visible metadata to preserve identity after YouTube replaces it with a generic placeholder.
Candidate scoring
Title similarity is combined with source channel and recording-version signals. Risk labels such as live, remix, cover, karaoke, remaster, lyrics, or sped-up reduce confidence.
No invented identity
A hidden count without a surviving video ID or prior metadata remains an anonymous gap and is not assigned a fabricated title.
Write boundary
What happens before and after a playlist write
Playlist writes begin only after the evidence model has finished. Each gate can stop the operation before YouTube is changed.
User policy check
Manual repair requires approval. Automatic repair requires an eligible plan and an explicit rule configured by the user.
Live-state validation
The playlist is re-read to confirm the damaged item, target position, and absence of a duplicate candidate.
Position-aware insertion
The chosen candidate is inserted at the intended manual-order position when YouTube permits it.
Result verification and undo
The resulting playlist item is verified and recorded with the original evidence so the change remains traceable and reversible.