Skip to content
Live demo

Smart Categorization

When you follow many feeds across different topics, new articles sometimes feel hard to prioritize. You have to open each one to understand its relevance.

Smart Categorization uses AI to automatically tag and sort articles based on your existing folder structure and reading history.

When a new article is fetched, the AI:

  1. Reads the title and summary
  2. Compares against your folder names and feed categories
  3. Assigns a relevance tag (“High Interest”, “Skim”, “Low Priority”)
  4. Suggests the best folder if the article came from an uncategorized feed

The model learns from your behavior. Articles you open and spend time reading nudge the model toward similar content in the future.

Each article gets a relevance score from 0–100:

ScoreLabelRecommended Action
80–100Must ReadAlways open these
60–79Worth ReadingOpen when you have time
40–59SkimCheck the summary only
0–39Low PrioritySkip or auto-archive

Articles below a configurable relevance threshold can be automatically marked as read without appearing in your unread count:

AI_AUTO_ARCHIVE_THRESHOLD=20 # articles with score < 20 are auto-archived

This dramatically reduces inbox overwhelm for busy feeds.

Smart Categorization works entirely locally when using Ollama. Your reading history and article content never leave your machine.