jev-ios-bridge
Claude Code plugin that runs tap scripts on iOS or Android and has Jev judge each checkpoint's on-screen claims for a pass/fail verdict.
Overview
jev-ios-bridge checks an iOS or Android app the way a person reads its screen. You write a script of taps, typing, swipes and waits with checkpoints such as 'The order total reads $5'; the bridge runs it on an iOS simulator through MobileBuildMCP, or on an Android emulator or phone through adb and a device agent. At each checkpoint Jev reads the screen text and judges the claims, producing one verdict (passed, failed or inconclusive) with screenshots and a log. An experimental 'do' step lets Jev choose the next action from those offered.
Scripts live in the app repository and work for SwiftUI, UIKit, Compose and Android Views. Claude Code writes and runs them through two plugin skills and steps in when Jev cannot decide, and saved scripts rerun from the terminal without Claude. The README itemises costs at jev-1.13.0 pricing: under $0.0002 per checkpoint and about $0.002 for a driven run of 13 decisions and 2 checkpoints. A list summary describes it as at a planning stage, though the README shows a v1.3.1 release.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Yes / no | Does the screen support this checkpoint claim? | — |
| Choice | Which offered action should the driven step take next? | — |
How it works
- Write a script of actions and checkpoints
- Bridge runs it on a simulator, emulator or phone
- At each checkpoint, Jev judges claims from screen text
- Return passed, failed or inconclusive with screenshots and logs
Strengths
- Very low cost per checkpoint
- Scripts rerun without an LLM
- Works across iOS and Android UI frameworks
Caveats
- Driven 'do' steps are experimental
- Relies on screen text, not visuals
Getting started
Install the Claude Code plugin and use /jev-ios-bridge:test-ios or /jev-ios-bridge:test-android.At a glance
- Type
- App
- Status
- beta
- Language
- TypeScript
- Runtime
- Claude Code plugin with MobileBuildMCP and adb
- Has code
- yes
Use cases and integrations
How curated lists describe it
Bridge that lets Claude Code verify an iOS app on a simulator, with TypeSafe's Jev choosing each step (planning stage).
hellogumbo/awesome-jev
Frequently asked questions
What does Jev decide?
Whether each checkpoint's claims hold on the current screen, and optionally the next action.
Does every run need Claude?
No, saved scripts rerun from the terminal without Claude.
What does a check cost?
Under $0.0002 per checkpoint, per the README.
Related projects
Fast iOS Simulator and Android navigator built on sim-use where Jev picks each tap instead of a frontier-model turn.
ProjectbetaLocal desktop pal that verifies what a coding agent really did, with optional Laya or Jev double-checks of unclear test results.
ProjectbetaTypeScriptDeterministic-first 'System 1.5' gate for coding agents: triages test failures, breaks doom loops and governs effort, with optional Jev.
CLIbetaPythonFork of mobile-mcp that adds OCR and lets Jev pick which recognised text element to tap on Android and iOS devices.
MCP serverbetaTypeScriptFlutter integration-test package where Jev picks the next registered UI action and judges whether the test goal is reached.
LibrarybetaDartPlain-English end-to-end tests for Android and iOS apps where Jev works out the taps and a lockfile replays every decision exactly.
AppbetaPython