iphone-use

Let AI agents drive a real iPhone, even apps with no API

Developer ToolsArtificial IntelligenceGitHubApple

▲ 70 votes1 commentsLaunched Oct 6, 2026

Daily #36Weekly #37

iphone-use gives an AI agent a real iPhone: the screen as text, tap, type and scroll over WebDriverAgent, with an honest answer for every action (applied, not sent, or unknown). Payment apps that block screenshots come back as a wireframe. A task done once becomes a flow that replays with no model. HTTP API, a 21-tool MCP server for Claude Code, and remote control from a browser or iOS app. Open source, MIT.

AI Analysis

📝 Summary

iphone-use empowers AI agents to control a real physical iPhone by representing the screen as text and enabling tap, type, and scroll actions via WebDriverAgent. It delivers honest feedback on every action (applied, not sent, or unknown) and converts screenshot-blocked apps like payment tools into wireframes. One-off tasks convert into reusable flows that replay without models. It offers an HTTP API, 21-tool MCP server optimized for Claude Code, plus remote control from browsers or iOS apps. Open-source under MIT, it solves the critical pain point of automating iOS apps lacking APIs, delivering high value for AI-driven mobile interactions.

📈 Market Timing

In 2025-2026, AI agent technology is maturing rapidly with surging demand for multimodal device control, aligning with trends like Anthropic's computer-use tools extending to mobile. User needs for reliable app automation beyond APIs are growing, supported by favorable AI innovation policies and economic investment in dev tools. This is an ideal window before widespread adoption. Excellent Timing.

✅ Feasibility

Leverages mature WebDriverAgent technology, keeping technical difficulty manageable for those with iOS/AI expertise. Open-source model lowers initial development costs, with good scalability potential for flows and API. Requires physical iPhones, raising hardware/operation costs and some Apple compliance risks, but no major supply chain issues. Overall High feasibility with proven implementation.

🎯 Target Market

Main segments: AI/ML developers, automation engineers, indie hackers and tech companies building agentic systems (ages 25-45, tech-savvy). Industries: Software development, AI services, fintech automation. Geographic: Global with concentration in US, Europe, East Asia. TAM for AI dev tools ~$20B+ by 2026; SAM for mobile agent infra ~$2B; SOM for iOS-specific ~$300M. Pain points: inability to control closed iOS apps without APIs. High willingness to pay for hosted/reliable versions.

⚔️ Competition

Medium. Direct competitors: 1. Appium (appium.io), 2. Maestro (maestro.mobile.dev), 3. Detox (wix.github.io/detox), 4. XCUITest (developer.apple.com), 5. emerging AI tools like MultiOn (multion.ai). Advantages: real iPhone focus with honest action feedback, reusable model-free flows, specific Claude integration and no-API handling. Disadvantages: physical device dependency vs simulator ease, limited enterprise support as open-source compared to commercial tools.

Upgrade Pro to unlock full AI analysis