expo-mcp-bridge

Let your AI agent use your phone app the way it uses a browser. It reads the screen, taps, types and fills forms, scrolls, takes screenshots, reads console errors and network requests, and presses volume / back / home — in your Expo app, running live on your phone.

What the agent sees

Instead of a screenshot (~1,000 tokens), the agent reads a compact outline of the visible screen (~70 tokens), with stable refs it can act on:

scroll [e1]
  button "icon:chevron-back" [e2]
  text "Welcome back"
  input "you@example.com" value="" [e3]
  input "••••••••" value="" (secure) [e4]
  button "icon:eye" [e5]
  button "Log In" [e7]

Every action returns only what changed:

Tapped.
Changes (+ added, - removed):
~ input "••••••••" value="" [e4]
+ button "icon:eye-off" [e9]
- button "icon:eye" [e5]

Highlights

• Real devices, zero native setup — Expo Go or a dev build, iPhone or Android, over Wi-Fi.
• 17 tools — read, act, debug and device keys. See the tools reference.
• Errors attached to actions — new JS errors and failed requests come back with the action that caused them.
• Skills included — init installs an agent skill and an AGENTS.md guide.
• Dev only — nothing runs in production builds.

Next

Install it, then connect your agent.