Kagi Small Web

Original Article ↗ Hacker News Discussion ↗

Concept and Nostalgia

  • Many see Kagi Small Web as reminiscent of StumbleUpon: a way to “stumble” across random interesting sites, especially personal blogs.
  • Several commenters find it fun and potentially addictive; others note there are already similar tools and directories.

Curation, Scope, and Criteria

  • Small Web’s index is manually curated and based on submitted sites stored in public text files (for blogs, comics, YouTube, HN, etc.).
  • Criteria: human-written, RSS feed, recent posts, blogs/webcomics only. Some argue this definition is too narrow and excludes classic “small web” gems, static info sites, and experimental or single-purpose pages.
  • Concerns that, even with human curation, the feed quickly fills with tech/AI/LLM/coding content and feels top‑heavy toward that subculture.

Language, Content Quality, and AI Slop

  • Multiple requests for language filtering; current setup is English‑centric and curation in other languages is seen as hard.
  • Several users are disappointed by AI‑generated or AI‑sounding posts in a feature marketed as “small web,” viewing this as against the ethos of highlighting real neighbors behind sites.
  • Others note that AI slop is now pervasive everywhere; Kagi is trying to derank it via heuristics but it’s incomplete.

UX and Technical Issues

  • Users want stable URLs for the first random page so they can return to it; currently a refresh often loses the page.
  • Some sites block embedding via X‑Frame‑Options, breaking the in‑frame experience; users end up opening pages in new tabs.
  • The “Next Post” is random; “Show Similar” uses semantic search. Some appreciate this, others want more obvious controls.

Broader Search and Web Context

  • Several comments zoom out: the open web is increasingly polluted, and all search engines are “searching in a pile of junk.”
  • Small Web is seen as an attempt to carve out a higher‑quality niche, but questions remain whether it can avoid being gamed as it grows.

Kagi as a Product and Strategy

  • Mixed views on Kagi overall: some report significantly better, cleaner results than Google and appreciate features like per‑site ranking and AI summaries; others find Kagi no better (or worse) and consider canceling.
  • A few feel Kagi is losing focus by spreading a small team across many projects (assistant, browser, Small Web) instead of concentrating on core search.