Cubox (Hoarder) vs NewsBlur
| Tagline | AI-powered self-hosted bookmarking and read-it-later with automatic tagging | Social RSS reader with story training and discussion features |
| Category | Feeds & Read-Later | Feeds & Read-Later |
| Replaces | Raindrop.io, Pocket, Instapaper | Feedly, Pocket, Instapaper |
| GitHub stars | 28k | 7.6k |
| Language | TypeScript | Python |
| License | AGPL-3.0 | MIT |
| Self-host difficulty | 2/5 Easy | 5/5 Advanced |
| Deploy options | Docker Compose One-Click | Docker Docker Compose Manual |
| Managed hosting | ||
| Last updated | 6 days ago | 8 days ago |
| View repo | View repo |
Where each falls short
The honest trade-offs — what you give up with each, versus the proprietary tools they replace.
Cubox (Hoarder)
- AI tagging requires Ollama or an OpenAI-compatible API; adds resource overhead
- RSS reader features are basic compared to dedicated feed readers like Miniflux
- Relatively young project; some rough edges in mobile app stability
NewsBlur
- Self-hosting is complex: requires MongoDB, PostgreSQL, Redis, Elasticsearch, and Celery workers
- Documentation for self-hosted setup is sparse compared to the hosted version
- Social blurblog network only active on the official hosted instance
- Story training model is less sophisticated than modern ML-based feed ranking in Feedly Pro
Bottom line
Choose Cubox (Hoarder) if you want the lower-effort setup; choose Cubox (Hoarder) for the larger community and ecosystem. Cubox (Hoarder) has seen more recent development. Open each guide below for deploy steps and the full feature gap.
Cubox (Hoarder)
AI-powered self-hosted bookmarking and read-it-later with automatic tagging