RSSMonster’s Philosophy

RSSMonster treats RSS not as a chronological inbox, but as an information ranking problem.

Modern RSS usage suffers from:

  • Repeated coverage of the same stories
  • Syndication noise
  • Uneven source quality
  • Ever-growing unread backlogs

RSSMonster addresses this by introducing signals, scores, and intent-driven views — while remaining fully self-hosted and under your control.


Articles

An article is a single item fetched from an RSS or Atom feed.

In RSSMonster, articles are more than raw entries:

  • They are analyzed for quality
  • Compared against other articles for uniqueness
  • Ranked relative to time, trust, and engagement

Articles remain immutable: RSSMonster never rewrites or alters feed content.


Clusters (Stories)

RSSMonster groups similar articles into clusters using semantic similarity.

A cluster represents:

  • One story
  • One topic
  • One news event covered by multiple sources

Why clustering matters

Clustering allows RSSMonster to:

  • Collapse duplicate coverage
  • Reduce noise from syndication
  • Compare originality between sources
  • Rank stories instead of posts

Reading one article contributes to reading the entire cluster.


Quality

Each article receives a quality score between 0.0 and 1.0.

Quality reflects:

  • Writing structure and clarity
  • Promotional language detection
  • Sentiment neutrality
  • Content richness

Quality is not a value judgment — it is a signal used for:

  • Ranking
  • Filtering
  • Automation rules

Uniqueness

Uniqueness measures how much new information an article adds relative to others in the same cluster.

Articles rank higher when they:

  • Provide original reporting
  • Add context or analysis
  • Are not near-duplicates

Syndicated copies and rewrites score lower, even if published later.


Freshness

Freshness measures how recent an article is, normalized over time.

Freshness:

  • Decays smoothly
  • Never drops to zero instantly
  • Is balanced against quality and trust

This prevents “latest wins” behavior while still surfacing breaking news.


Feed Trust

Feed trust is a long-term score (0.0 – 1.0) earned by each source.

Trust reflects:

  • Originality of published articles
  • Average quality
  • User engagement (reads, clicks, stars)
  • Publishing consistency

Trust is earned over time, not configured manually.

High-trust feeds influence ranking more, but never silence others.

Adaptive feed scheduling

RSSMonster estimates publisher activity from valid, non-future entry timestamps, including entries that were already known, deduplicated, or filtered locally. It sorts and deduplicates the timestamps, derives adjacent publication intervals, and feeds their median into a bounded EWMA. A first feed snapshot records only the newest publication time; it does not infer cadence from a potentially large historical backlog.

This intentionally differs from CommaFeed’s arithmetic average over one feed snapshot. RSSMonster uses publisher evidence across observations and a median sample so duplicates, ordering, and one long gap have less influence. Adaptive fetch intervals remain bounded between five minutes and four hours.


Engagement Signals

RSSMonster learns from how you actually read.

Engagement signals include:

  • Reading time
  • Scroll-based mark-as-read
  • Clicks
  • Starred articles

These signals are aggregated to:

  • Improve feed trust
  • Stabilize ranking
  • Reduce noise over time

RSSMonster does not track behavior externally.


Importance

Importance is a runtime score that determines article ranking.

It combines:

  • Freshness
  • Quality
  • Uniqueness
  • Feed trust

Importance answers:

“How likely is this article worth my attention right now?”

It is recalculated dynamically, not stored permanently.


Smart Folders

Smart Folders are declarative, dynamic views built from search expressions.

They:

  • Replace static folders
  • Encode intent
  • Update automatically

Example:

@today unread:true sort:recommended

Smart Folders allow you to define what matters — not just what arrived.

Automation & Actions

RSSMonster supports automated actions using regular expressions and signals.

Actions can:

  • Star articles
  • Mark articles as read
  • Delete low-quality content
  • Flag advertisements

Automation works with scoring, not against it.
Instead of relying on fragile keyword rules alone, actions can be combined with quality, freshness, and trust signals to reduce noise safely.


Transparency by Design

RSSMonster is built on transparent principles:

  • No black-box algorithms
  • No external tracking
  • No advertising incentives
  • No forced personalization

Every decision can be:

  • Inspected
  • Filtered
  • Overridden

You decide how much automation and ranking you want — nothing is hidden or imposed.


Who These Concepts Are For

These concepts matter most if you:

  • Follow many overlapping sources
  • Care about signal over volume
  • Want control over ranking
  • Prefer explainable systems

If you only follow a few feeds, RSSMonster will still work — but its strengths shine at scale.


Summary

RSSMonster introduces:

  • Clusters instead of repetition
  • Signals instead of guesses
  • Ranking instead of inboxes
  • Control instead of magic

It is not a replacement for RSS.

It is RSS — taken seriously.


This site uses Just the Docs, a documentation theme for Jekyll.