Blockbusterr v2.0.0: self-hosted media discovery with rules, previews, and delivery limits

Hi everyone! I'm the developer of Blockbusterr, an MIT-licensed, self-hosted tool for deciding which movies and shows reach your media library. Today I'm announcing v2.0.0! It sits between discovery sources such as TMDB, Simkl, Trakt, and MDBList and your delivery tools: Radarr, Sonarr, Jellyseerr, or Seerr. It evaluates candidates against your rules instead of importing an entire trending list or watchlist. Your existing stack handles downloading and playback. The main changes in v2: Reusable movie/show rules, with job-specific copies and universal title exceptions. Job previews that show accepted, rejected, and already-present titles before anything is sent downstream. Per-job and rolling delivery limits, repeat handling, and optional ranked selection across jobs. Supported lists/watchlists alongside trending and other discovery jobs. Activity Entries and Job Runs explaining what happened to each title. Automatic migration from v1, plus separate shareable configuration exports and private backups. It runs as a single Docker container with SQLite. You need one discovery provider and one delivery target; Trakt is optional. Direct public Letterboxd scraping is available as an experimental opt-in integration. For example, you could follow a trending movie feed but only deliver a few titles that meet your language, year, rating, and genre rules. Preview lets you check those choices before turning on automation. Existing users: latest moves from v1 to v2. Back up the full data directory and follow the upgrade guide; start with dry-run while checking migrated jobs. Keep Blockbusterr on a trusted network or behind authenticated access. Source and screenshots Quick start Release notes Upgrade guide I'd appreciate feedback on the preview/rules workflow and reports from v1 upgrades. If something goes wrong, please include the version, job type/provider, expected behavior, and redacted logs in a GitHub issue.