Every article I read has words I half-know. Not completely foreign — half-known, the kind where I could probably reconstruct the meaning from context but I'd be guessing. I kept making the same trade-off: open a new tab and lose my place, or skip past the term and stay shallow.
I wanted neither. I wanted the explanation to appear where I was already looking.
So I built rabbitholes — a Chrome extension that renders an inline explanation next to your cursor when you highlight any text. The tooltip sits in a shadow DOM so it doesn't interfere with the host page's styles or layout. Requests go directly from your browser to api.anthropic.com; there's no intermediary server, no telemetry, nothing phoning home.
The part I actually use most: any word in the explanation is itself clickable. Drag across words to select a phrase. The extension tracks how many hops deep you've gone — I've hit eight reading about Byzantine monetary policy before catching myself.
Two other things worth knowing:
- A globe icon re-runs the query enriched with Brave Search results, with source chips you can click through. Useful when the first answer is too abstract and you want a grounding example.
- Every answer ends with two suggested rabbit-hole threads — the model's read on where the most interesting branches are from here.
The free-form follow-up (pencil icon) inherits the current context as background, so you can ask a specific question without re-explaining what you were reading about.
Manifest V3. API key stored in chrome.storage.sync — encrypted, never leaves the browser.
// The shadow DOM attachment — keeps tooltip styles
// from leaking into or inheriting from the host page
const shadow = container.attachShadow({ mode: 'open' });
const styleEl = document.createElement('style');
styleEl.textContent = tooltipStyles;
shadow.appendChild(styleEl);
shadow.appendChild(tooltipRoot);
Top comments (0)