Show HN: A 15 MB self-hosted Netflix and Spotify for your media server
A new open source tool aims to replace the usual six-application “*arr” stack (Sonarr, Radarr, Lidarr, Bazarr, Prowlarr, Seerr) with a single 15 MB Go-based service that gives a Netflix- and Spotify-style interface for managing downloads into Plex or Jellyfin. Commenters are intrigued by the simplicity, small footprint, and AI-assisted development, but raise concerns about relying on a central metadata server, the project’s youth and security profile, and its positioning as an explicit piracy workflow. Others question whether consolidating specialized tools is wise, how long such a vibe-coded project will be maintained, and what legal or ethical risks contributors might face.
Project goals & architecture
- Tool aims to unify the “arr” acquisition stack (Sonarr, Radarr, Lidarr, Bazarr, Prowlarr, Overseerr) into a single app that feels like Netflix/Spotify for browsing, but delegates playback to Plex/Jellyfin/Emby.
- Single Go binary with Svelte UI and SQLite; Docker image is ~15 MB, with no OS or shell in the container.
- Integrates with existing downloaders (qBittorrent, SABnzbd) and media servers; can import Sonarr/Radarr configs and speak their APIs so existing tools keep working.
- Runs on amd64, arm64, armv7; AGPL-3 licensed; uses a separate metadata server for posters, charts, and search to avoid user API key setup.
Size, stack, and “what was removed”
- Discussion emphasizes that most of the size reduction vs. the arr stack comes from:
- Avoiding bundled OS layers and runtimes (.NET, Python, Node) in multiple containers.
- Consolidating overlapping logic across the different arr apps.
- Some skepticism that a small, “vibe coded” project can match mature tools without hidden compromises; others argue image size is mostly bloat/runtime and not a proxy for feature parity.
Metadata centralization & self-hosting concerns
- Several commenters are uneasy about reliance on the project’s hosted metadata server, seeing it as a central point of failure and contrary to self-hosting ideals.
- Others note the existing arr apps already rely on their own central metadata APIs.
- Roadmap mentions future support for user-supplied API keys/providers.
AI involvement in development and features
- Large portions of the code were reportedly generated by an AI assistant, with human review and testing.
- Some are impressed by what AI-enabled solo development can achieve; others distrust quality and doubt “full review” claims given commit sizes.
- Debate over adding AI/MCP features: proposed uses include natural-language search (“that Christmas movie with the kid alone”), subtitle translation, SMS/voice request bots, and automated library maintenance. Others see these as unnecessary complexity or wasteful.
Maturity, security, and stability
- Concern that the repo is very new publicly (hours old) with many rapid 1.0.x releases, raising stability and security questions.
- Clarification that it was private and used/tested for months, with vulnerability testing before public release; first public week described as “beta-like.”
- Some argue releasing early is fine for open review; others prefer to wait until it “cooks” longer.
Feature gaps and wishlist
- Music support and some integrations (Plex/Trakt watchlists, more subtitle providers, “bring your own” metadata keys) are incomplete or planned.
- Users request AI-based subtitle translation, alternative subtitle providers, better Blu-ray ripping workflows, and easier agent-based or scripted setup.