Add browser agent tools to let AI help users on your page
Adds WebMCP tools so browser agents can read and act on your page, like marking tasks or filtering a list.
When to use it
Your task app needs AI to help users mark tasks complete and filter by status without building a backend server.
Why it is worth it
Agents help users interact with your app without needing a separate server, using their signed-in session.
The prompt
I have a task management app with a task list. Users can see all tasks, mark tasks as complete with a checkbox, and filter by status (all, active, completed). Add browser agent tools so an AI agent can: 1) read the current list of tasks and their status, 2) mark a task as complete by clicking its checkbox, 3) filter the list to show only active or completed tasks. Start with read and update tools only, not delete tools. Report which tools you added and confirm the agent can read tasks and mark one as complete.
Replace the task app with your own app and its main actions users need help with.
What you get
Your app has WebMCP tools letting browser agents read and take actions on your page without needing a backend.
From docs.lovable.dev/changelog#let-browser-agents-use-your-app
