ShawnPana / phone-harness

A unified harness that lets any AI agent see, tap, type, and read results on real iPhones and Androids—no jailbreak, no Xcode, no phone modifications needed.

活跃维护 MIT Python Tracked
4.0k 400 2 天前
CIPyPI
agent ai automation developer-tools mobile-ai mobile-development

星标趋势

数据积累中,暂无足够数据生成趋势图

AI 分析

项目摘要

phone-harness is a Python tool that lets AI agents (Claude Code, Codex, or any agent) control real mobile devices—iPhone via Mac Mirroring, Android via ADB, or cloud-hosted phones—through a unified Python API for screen reading, tapping, typing, and app navigation.

为什么值得关注

It bridges the gap between coding agents and real mobile devices without requiring jailbreaks, Xcode, or anything installed on the phone, enabling AI agents to perform end-to-end mobile tasks like booking rides or testing apps through natural language.

优势

  • Unified Python API across iPhone (Mirroring), Android (ADB), and cloud-hosted phones
  • Zero phone modification required — no jailbreak, no Xcode, nothing installed on device
  • Agent-native design with intuitive helpers like tap_text(), type_text(), screen(), open_app()
  • Cloud phone offering with real numbers, 2FA support, and unlimited devices for testing without hardware
  • Very active development with 8 releases in 6 months and strong community adoption (3.6k stars)

局限性

  • Small team of only 3 contributors raises long-term sustainability concerns
  • iPhone support limited to macOS via iPhone Mirroring, not cross-platform
  • Complex setup chain requiring Accessibility and Screen Recording permissions on Mac, or Android developer options
  • No Docker support for containerized deployments
  • Documentation appears truncated in README with incomplete usage reference

使用场景

  • AI-driven end-to-end mobile app testing and QA automation
  • Automating mobile workflows (ride booking, messaging, form filling) through natural language agent commands
  • Cross-platform mobile development testing using cloud phones without physical devices
  • Mobile app prototyping and demo creation via coding agents
目标用户: AI developers, mobile developers, and QA engineers who want to extend their coding agents' capabilities to real mobile devices
学习曲线: 中
分析模型:LongCat-2.0 | 分析时间:10 天前