Personal RSS news reader with intelligence training, original site view, and social sharing
MIT
- Python
- JavaScript
- HTML

About NewsBlur
NewsBlur is a personal news reader for following RSS feeds, reading full stories, and organizing articles from one place. It shows the original site while adding filtering tools so you can keep up with sources and train the reader to highlight what you want and hide what you do not.
It supports real-time RSS, story training, full text search, saved searches, story tagging, multiple layouts, dark mode, and reading email newsletters. It also includes shared stories, blurblog privacy controls, YouTube channels, third-party app support, IFTTT integration, and an MCP server for connecting AI agents to feeds, stories, and classifiers.
NewsBlur runs as a Django application with PostgreSQL, MongoDB, Redis, optional Elasticsearch, Celery, and Node.js services. It offers free use at newsblur.com up to 64 sites, premium plans, native iOS, macOS, and Android apps, and a self-hosted instance for users who want to run it with control over their data.
Key features
- Real-time RSS feeds pushed as stories arrive
- Train story filters to highlight or hide items
- Original site view and full text expansion
- Shared stories, blurblog privacy, and story tagging
- MCP server for AI agents and classifiers
Details
- First released
- 2009
- Platforms
- Web · Android · iOS · macOS
- Self-hosting
- Docker · Docker Compose
- Storage
- PostgreSQL · MongoDB · Redis
- Search
- Elasticsearch optional
- Pricing
- Free up to 64 sites; premium plans
