Sprint 1 — Auto-refresh after download:
- New [[mediaserver]] TOML section with kind/url/token/sections
- mediaserver.Refresh() fans out to Plex (partial via section ID auto-mapping
from file path prefix) and Jellyfin/Emby (full library scan)
- Manager.OnFinalized callback wired in daemon to trigger refresh after
organize() completes — keeps engine package free of mediaserver dep
- New unarr mediaserver {setup,list,remove,test} commands
- unarr init wizard offers to configure refresh when a server is detected
Sprint 2 — .strm instant mode (cloud + agent):
- Mode strm-to-library handled in daemon dispatch: writes a one-line .strm
file pointing to the cloud-resolved debrid HTTPS URL, then triggers refresh
- engine.WriteStrm + StrmDestForTask mirror organize()'s naming so Plex/Jellyfin
see the expected folder structure (Movies/Title (Year)/, TV Shows/Show/Season XX/)
- Atomic write (temp + rename) so partial files never get indexed
- Reports completed/failed status to the cloud via existing agent client
- Migration wizard from Sonarr/Radarr/Prowlarr (unarr migrate) [pre-beta]
- Auto-detect instances via Docker, config files, port scan, Prowlarr
- Import wanted list (monitored+missing movies/series)
- Import download history and blocklist to avoid re-downloading
- Extract debrid tokens from *arr download clients
- Quality profile mapping to preferred_quality config
- DISTINCT ON PostgreSQL query for optimal torrent selection
- JSON export with --dry-run --json (text to stderr, JSON to stdout)
- Media server detection (Plex/Jellyfin/Emby) in unarr init
- Detects library paths and offers them as download directory options
- Debrid auto-configuration in unarr init
- Scans *arr instances for debrid tokens
- Validates and saves via API if user confirms
- New preferred_quality setting in config (2160p/1080p/720p)
- Library scan command (unarr scan) with ffprobe metadata extraction
