Lenss

Right-click an AI reply. Notice what it is doing — not just what it says.

Lenss is a small, calm instrument for the assistant text in your AI conversations which can help you improve your communication by helping you understand the response and where it leads. Lenss is built around one decision filter: does this preserve the perceptual snap? The snap is the moment of "oh, I didn't notice that before" — and everything that competes with it has been deliberately stripped out. Highlight a passage in a ChatGPT or Claude reply. Trigger Lenss (right-click → "Lenss it", or ⌘⇧L / Ctrl+Shift+L). A 420-pixel overlay appears next to the selection. One or two short notices fade in — observations about what the writing is doing, in a single sentence each, anchored to a verbatim quote from the passage. Sometimes a single quieter line follows, naming a different trajectory the conversation could take. That's the whole extension. No accounts. No login. No popup, no settings page, no dashboard. No reading until you ask. No background monitoring. No history, no sync, no notifications, no badges. No coaching, no rewriting, no "try saying X instead." No third signal, no "show more," no recursive depth. Lenss does not generate text on your behalf. It is a reflective layer above AI generation, never the generation itself. The companion that drives the reveals is a small Cloudflare Worker maintained alongside this extension. The full privacy policy lives at lenss.one/privacy. The web companion at lenss.one offers a slower, deeper read of the same instrument — paste a passage and see more dimensions of what it's doing. The extension is the fast, ambient counterpart: same instrument, different cognitive mode.