Browse the web, create images and Deep mode

Hey there!

The AI Assistant’s composer has a new + menu. Attaching files is still there, and next to it are five ways to tell the Assistant what kind of answer you want — no prompt engineering needed.

Browse the web

Pick Browse the web and the Assistant searches the web before it answers. Paste a link and it reads the page too, so the reply rests on what it actually found, not on memory.

Create image

Describe a picture and the Assistant draws it. It arrives as a card under the reply, with Insert into document and Save to project. Right in the editor, the slash menu has Create image as well: describe it where the picture should go, and it’s drawn in place.

Deep mode

Deep mode brings in a second model, from another family, to review what the first one wrote and changed before the turn reaches you. It’s the mode for a long rewrite or a careful edit. Turn it on from the + menu or the Deep mode button next to the send arrow, and the Assistant works longer, with twice as many steps.

Polish & finalize

Pick Polish & finalize when a document is ready to go out. The Assistant takes it to its final version for export — wording, flow, structure and formatting, with no placeholders left — so what you send or print is finished.

Agents review

Agents review hands your document to several agents at once, each reading it the way a different reviewer would. They come back with suggestions to improve it — what’s unclear, what’s missing, what’s weak — and you choose which to take.

Browse the web and Polish & finalize are on every plan. Pictures come with Plus and Pro, and Deep mode and Agents review are part of Pro.

Also in this update

Updates now download and install in the background. When one is ready, a Restart to update button appears in the title bar — restart when it suits you, and you’re on the new version.

Try the new menu and tell me what you build with it. Cheers, Vagan