— the shortest path through what’s known

Zetlyn plans the reading order for a question, so every page you open adds something the one before it did not.

WebNewsStandards
What are you trying to find out?
Search infrastructure for agents

Read the web in the right order.

Zetlyn turns a result list into a reading route. Copies, sibling domains and repeated wording move aside, so the next page is more likely to add a new source, a new angle or a useful citation path.

For an LLM or an agent this is not a nicety. Every page opened costs tokens, money and latency, and a page that repeats the last one costs all three for nothing — so a shorter route to the same answer means less context spent, a smaller bill and a result sooner.

Order results by what they add Collapse copies and sibling domains Follow citations and related subjects Explain every relation used
Typical result list Same topic, repeated
1
Original technical guide
publisher-a.example
source
2
The same guide, republished
mirror-one.example
duplicate
3
A sibling-domain rewrite
guides-a.example
same owner
4
Independent explanation
publisher-b.example
new source
5
Syndicated copy
mirror-two.example
near duplicate
Zetlyn removes the repetition
A cleaner reading route
1
Original technical guide
publisher-a.example
seed
2
Independent explanation
publisher-b.example · was #4
new voice
3
Benchmarks and limits
lab-c.example · was #23
new ground
4
Who relies on the guide
docs-d.example · was #41
cites
5
A second lab, same benchmark
lab-e.example · was #58
corroborates
Three repeated results were skipped. The route keeps the relation behind every step visible, so an agent can inspect why it was chosen.