What is agent-device used for?

agent-device lets AI coding agents inspect, control and verify apps on mobile and other platforms through a CLI, MCP server or typed Node.js API. It is a good fit for letting coding agents verify mobile UI changes, automated end-to-end checks on simulators and cross-platform app debugging by agents.

Is agent-device open source?

Yes. agent-device is open source under the MIT licence. Its source code is on GitHub at callstack/agent-device and is written mainly in TypeScript.

Is agent-device free?

Yes. agent-device is open source, so the software itself is free to use.

What are some alternatives to agent-device?

Similar open-source tools in the Testing & QA category include Detox, Lamda and Playwright. SaaS products in the same category include Firebase Test Lab, Agent Checker and AlwaysQA.

Is agent-device actively maintained?

Yes. The most recent commit to agent-device was on 2 October 2026, and the latest release is v0.21.18, published on 30 September 2026. The project has 4.9k stars on GitHub.

agent-device: open-source testing & QA

About agent-device

agent-device is a tool that gives AI coding agents a live feedback loop with the apps they are changing. Through a command-line interface, a built-in MCP server or a typed Node.js API, an agent can launch an app, inspect it, interact with it and check that its code changes actually work. It comes from Callstack, known for its React Native work.

Supported targets include iOS, Android and HarmonyOS on simulators, emulators and real devices, plus tvOS, Android TV, Amazon Vega OS, web, macOS and Linux. Instead of reasoning only over screenshots, agents read token-efficient accessibility snapshots, act through element references and selectors, and save evidence for review. Repository topics reference XCUITest, ADB, Expo, React Native and Flutter, as well as end-to-end testing and performance work.

The project is MIT-licensed and written in TypeScript. It runs on your development machine alongside simulators and your chosen coding agent. It suits mobile developers who want agents to verify UI changes rather than just edit code.

Key features

CLI, MCP server and Node.js API
iOS, Android and HarmonyOS automation
TV, web, macOS and Linux targets
Accessibility snapshots instead of screenshots
Actions through refs and selectors
Saved evidence for review

Good fit for

→Letting coding agents verify mobile UI changes
→Automated end-to-end checks on simulators
→Cross-platform app debugging by agents

Built with: TypeScript; React Native; Flutter
Tags: mobile-testing; ai-agents; mcp; automation; react-native; ios; android; e2e-testing

agent-device: questions and answers

What is agent-device used for?: agent-device lets AI coding agents inspect, control and verify apps on mobile and other platforms through a CLI, MCP server or typed Node.js API. It is a good fit for letting coding agents verify mobile UI changes, automated end-to-end checks on simulators and cross-platform app debugging by agents.
Is agent-device open source?: Yes. agent-device is open source under the MIT licence. Its source code is on GitHub at callstack/agent-device and is written mainly in TypeScript.
Is agent-device free?: Yes. agent-device is open source, so the software itself is free to use.
What are some alternatives to agent-device?: Similar open-source tools in the Testing & QA category include Detox, Lamda and Playwright. SaaS products in the same category include Firebase Test Lab, Agent Checker and AlwaysQA.
Is agent-device actively maintained?: Yes. The most recent commit to agent-device was on 2 October 2026, and the latest release is v0.21.18, published on 30 September 2026. The project has 4.9k stars on GitHub.