Observed arrival · 2026-09-23
T-Search Lab makes search-agent behavior inspectable
A research demo for running live web searches and examining search-agent traces, evidence memory, and behavior diagnostics.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- Researchers evaluating web-search agents and retrieval behavior
- Worth noticing
- Its diversity metric uses lowercased word-set Jaccard overlap, explicitly excluding stemming and semantic similarity.
Field notes
The interface separates a live query from four published traces and offers replay and trace export. A run is instrumented across rounds, searches, sources, context growth, timing, and agent actions; the page explains that live retrieval returns titles, URLs, and snippets rather than fetching full pages. Its lexical-diversity chart uses lowercased word-set Jaccard overlap, with no stemming or semantic comparison.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
✦PrettyNotable craft visible
◎NicheUnusually specific use
ƒJavaScriptBrowser-side code central
One card from the complete issue