Plembfin

Use Export on the old installation, then import the downloaded JSON file here. The backup includes watch data, settings, and cache metadata, but not stored artwork files. Treat it as sensitive because saved media-server credentials and API keys may be included.

Reads every record in your Firestore watchHistory collection and pushes the watched or unwatched state to each enabled media server. Use this after setting up a new Plex, Emby, or Jellyfin server, after a library rebuild that wiped playstates, or after restoring a server from backup. The sync respects your enabled and disabled flags from Apps settings. Progress is streamed to the log in real time and each server is processed independently, so a failure on one does not block the others.

Accepts the CSV and JSON files from a Trakt data export (unzipped). Files are parsed entirely in your browser and sent to /api/import in batches of 100 records. Each imported entry is written to watchHistory with the source field set to trakt_import, so it is always distinguishable from live webhook records. Duplicate detection is performed on the server — records that already exist are skipped, not overwritten. After importing, use Backfill Trakt Imports below to fetch poster artwork for the new rows.

Runs a live diagnostic against every backend dependency the dashboard relies on. Checks include: Firestore read and write access, the Firebase Functions API reachability, and a connection test against each configured Plex, Emby, and Jellyfin server using your saved credentials. Results show the latency and status of each probe individually so you can pinpoint which component is failing. Run this before making configuration changes, after a deployment, or when diagnosing unexpected sync failures or missing data.

Queries TMDB for each record in watchHistory that may have an incorrect type value — for example, a movie that was stored as a TV episode because the webhook payload was ambiguous. When TMDB returns a different type, the stored record is updated and the poster URL cache is refreshed. Only rows where the lookup disagrees with the stored type are modified, so it is safe to run at any time without risk of data loss. The repair log streams each corrected row in real time.

Scans every watchHistory row where the source is trakt_import and the poster URL is empty, then looks up artwork from TMDB and writes the URL back to Firestore. Runs in configurable batches with a per-call rate-limit delay to stay within TMDB's free-tier request quota. Only imported rows are touched — live webhook records are never modified. Run this once after a Trakt import, or again if poster artwork is still missing after a previous partial run.

Fetches full TMDB detail pages — cast and crew, trailers, reviews, and related titles — for every movie and TV show currently in your watchHistory, writes the results into the tmdbMetadataCache Firestore collection, downloads each poster and backdrop into Storage, and stamps the canonical poster back onto every watch record. This brings existing media to full parity with newly-added media: both detail pages and grid thumbnails are instant rather than waiting for a live API call. Run this after a large import or periodically. The cache is also updated automatically when a title is added or first viewed.

Scans every document in watchHistory, groups them by mediaKey (the canonical title and ID combination), and deletes all but the most recently logged entry for each unique item. Duplicates can accumulate when webhooks fire multiple times for a single watch event, after re-importing a Trakt export that partially overlaps existing records, or after a database migration that ran twice. No watched data is lost — only the older copies of an already-recorded item are removed. Safe to run at any time.