Your local AI coding agent can now drive the live AppFunnel editor. Open a funnel, get a pairing code, and connect Claude Code, Codex, or Cursor to that exact tab. From there the agent authors the funnel the way you would: reading your files, making edits, watching the preview rebuild, and taking screenshots to check its own work.
Use cases
- Build by conversation - tell your agent to add a paywall or restyle a page and watch it happen in the open editor
- A real feedback loop - the agent edits, sees the preview recompile, and screenshots the result to verify before moving on
- Your session, your rules - every change runs under your own login, so the agent can only do what you can
Impact
Coding agents are excellent at writing code but blind to whether it actually looks and works right. Wiring one into the live editor closes that gap: it edits real funnel source, sees the same live preview you do, and confirms the outcome with its own eyes, all inside the tab you already have open. Pairing is a code you approve, the connection is scoped to that one editor, and nothing runs outside your session. It turns "describe the change and hope" into an agent that builds, watches, and checks, with you in the loop the whole way.
