Ask about the page you’re already on.

Lensia answers questions about the page you have open in Chrome. Ask the sidebar about the whole page, or select one part of the screen, like a chart or a dense paragraph, and ask about just that.

In Ask for approval mode, the default, Task reads, scrolls and searches the page, waits for it to load, switches between tabs it can use, and takes screenshots. Anything that acts, like clicking, typing, opening a tab or going to another page, waits for your OK. Most actions can be allowed once or for that site.

Full access skips those prompts, after a warning each time you turn it on. Either way, it still asks before a payment, order or trade it recognizes, a pop-up confirmation, leaving unsaved changes, or submitting on a site you didn’t ask it to use. It never types into password fields or fields a site marks as card details, so enter payment details yourself. You can stop it from the page or the sidebar.

Free gives you limited usage, enough to try Lensia on a few real questions, with a Google sign-in and no card. It includes the floating panel, the sidebar with history and bookmarks, Lensia Max and the Essential models at Instant effort, page summaries, and PDFs up to 2 MB per message.

Web search, Fetch URL, the Advanced models, higher effort and Personalization start on Basic. Task and the Frontier models are on Pro. When the allowance runs out, wait for the reset, upgrade, or switch to a free pool of open-source models while that’s offered.

Lensia Max is the default model choice. In the sidebar it picks a model for each message from the models in the Essential and Advanced tiers your plan and effort allow, and labels the reply with the model that answered. It never picks a Frontier model; on Pro you choose those yourself.

In the floating panel at Instant effort, it sends your request straight to Claude Haiku 5.5, because routing would add a delay, and the answer is labelled Lensia Max. Either way, it costs what the model that answered costs.