agent-device
Mobile app automation and verification for AI coding agents. CLI, MCP server, and typed Node.js API for iOS, Android, HarmonyOS, TV, web, macOS, and Linux.
- 5.1k stars
- 345 forks
- 83 open issues
- Language: TypeScript
- License: MIT
- updated
Install
git clone https://github.com/callstack/agent-device && cp agent-device/*.md ~/.claude/agents/- Clone the repository and copy the agent .md definitions into ~/.claude/agents (or .claude/agents inside a project).
- Start a new Claude Code session to load the agents.
- Delegate work to them with the Task/Agent tool or by name.
Editor's note
Agent-device is a TypeScript CLI that gives AI coding agents direct control over iOS simulators, Android emulators, physical devices, tvOS, Android TV, macOS, and Linux desktop targets to verify what a mobile or desktop app actually does at runtime. It integrates with Claude Code (and other agent terminals like Cursor and Codex) by exposing a set of commands that open app sessions, take compact accessibility snapshots with short element refs such as @e1 and @e2, perform taps, fills, scrolls, and gestures, and collect debugging evidence including screenshots, video recordings, network traffic captures, CPU and memory performance samples, and React render profiles. Apps built with Expo, Flutter, and React Native are supported alongside fully native iOS and Android projects. A standout detail is the replay system: interactions can be recorded into replayable .ad scripts that run in CI pipelines, turning exploratory agent sessions into repeatable end-to-end checks. Mobile developers, React Native engineers, and teams running agentic development loops benefit most.
4 items in this repository
Automates Apple-platform apps (iOS, tvOS, macOS), Android devices, and Amazon Vega OS TV apps in Vega Virtual Devices. Use when navigating apps, taking snapshots/screenshots where supported, driving TV remotes, tapping, typing, scrolling, extracting UI info, collecting evidence, or planning agent-device CLI commands.
Verify and debug native, React Native, Expo, or Flutter apps on an Android Emulator with agent-device. Use when an agent needs to launch an app, inspect its live UI, tap, type, scroll, validate a code change, collect failure evidence, or reproduce a workflow on an Android virtual device.
Systematically explore and test a mobile app on iOS/Android with agent-device to find bugs, UX issues, and other problems. Use when asked to dogfood, QA, exploratory test, find issues, bug hunt, or test this app on mobile.
Verify and debug native, React Native, Expo, or Flutter apps on an iOS Simulator with agent-device. Use when an agent needs to launch an app, inspect its live UI, tap, type, scroll, validate a code change, collect failure evidence, or reproduce a workflow on an iPhone or iPad Simulator.
Subagents overview
What people ask about agent-device
What is callstack/agent-device?
callstack/agent-device is subagents for the Claude AI ecosystem. Mobile app automation and verification for AI coding agents. CLI, MCP server, and typed Node.js API for iOS, Android, HarmonyOS, TV, web, macOS, and Linux. It has 5.1k GitHub stars and its last recorded update is dated 2026-10-23.
How do I install agent-device?
You can install agent-device by cloning the repository (https://github.com/callstack/agent-device) or following the README instructions on GitHub. ClaudeWave also provides quick install blocks on this page.
Is callstack/agent-device safe to use?
Our security agent has analyzed callstack/agent-device and assigned a Trust Score of 97/100 (tier: Verified). See the full breakdown of passed checks and flags on this page.
Who maintains callstack/agent-device?
callstack/agent-device is maintained by callstack. The last recorded GitHub activity is dated 2026-10-23, with 83 open issues.
Are there alternatives to agent-device?
Yes. On ClaudeWave you can browse similar subagents at /categories/agents, sorted by popularity or recent activity.