Every HTML card has two buttons: Edit opens the built-in visual editor, and AI hands the card to an agent. Both work on the same file, so you can switch between them.
Edit with AI
Click the AI button on a card to open the Board Agent focused on that card. Describe the change in plain language (“make the headline bigger and the background darker”) and the agent rewrites the card’s HTML in place — internally through its update_element_html action — and the canvas preview updates.
You do not have to start from a finished card: ask the Board Agent to create one and it adds the element and writes the HTML for you.
A terminal agent (such as GLAI) can also edit the underlying HTML/Markdown files directly when you ask it to.
Edit in the visual editor
Click the Edit button on a card (or open a Markdown element) to launch the visual editor and change it by hand without writing code. It opens full-screen over the canvas.
- HTML — edit the rendered card visually; changes are saved back to the file.
- Markdown — a dedicated Markdown toolbar gives you headings, bold/italic, lists, links, and the usual formatting, with a live preview.
Edits autosave back to the element’s file, so the canvas preview and any later render reflect your latest version. Close the editor to return to the board.
Using both together
A common flow is to let the Board Agent rough out a card, then fine-tune details in the visual editor — or author the structure yourself and ask the agent to polish the styling. Because both edit the same file, your changes compound instead of conflicting.
Leave a Reply