Reading a dense article about monetary policy last week, I hit "Triffin dilemma" and did what I always do: opened a new tab, read half of Wikipedia, forgot what I was originally reading, and spent twenty minutes somewhere else entirely. The context switch isn't a minor annoyance — it compounds. By the time I return, the thread I was following has gone cold.
So I built rabbitholes: a Chrome extension where highlighting any text renders an explanation in a shadow-DOM tooltip next to your cursor. You never leave the page.
The shadow DOM detail matters. A lot of overlay extensions inject markup directly into the host document, which breaks host styles or gets clobbered by them. Shadow DOM keeps the tooltip isolated — no style pollution in either direction.
The part I actually use most: every word in the explanation is itself clickable. Highlight "Triffin dilemma," get an explanation that mentions "Bretton Woods," click that phrase, go deeper. There's a hop counter so you can see how far you've traveled from the original article. Go deep enough on almost anything and you hit philosophy — the trail is shareable.
Two other features worth naming:
Globe icon → re-runs the query enriched with Brave Search results
Pencil icon → free-form follow-up that inherits the current context
The search-enriched mode returns source chips you can click. It's useful when you want primary sources rather than an explanation.
On the privacy side: zero telemetry, no intermediary server. Requests go directly from your browser to api.anthropic.com and api.search.brave.com. Your Anthropic API key lives in chrome.storage.sync — encrypted, never leaves the browser. Manifest V3.
The mental model I was chasing: every concept in every article I read should be one highlight away from becoming clear, without breaking the read. That's what it does.
Top comments (0)