Ever wished you could tell your Mac “open Safari and bookmark this page” and watch it happen automatically? Peekaboo makes that reality by combining pixel-perfect screen capture with AI vision models that understand your interface. It’s like having a digital assistant that can actually see and interact with your apps.
What sets this apart is the dual deployment: use it as a CLI tool for scripting, or plug it into Claude Desktop/Cursor as an MCP server for conversational automation. The AI can identify UI elements by description (“click the blue Save button”), capture multi-screen setups at Retina resolution, and even discover menu structures without clicking around. It supports everything from GPT-4 to local Ollama models, so you control where your screenshots go.
With 1600+ stars and active development, this is becoming the go-to solution for Mac automation that doesn’t break when Apple changes the UI. Perfect for QA engineers, power users, and anyone tired of repetitive clicking. The beta v3 adds agent workflows that chain multiple actions together - it’s getting scary good at understanding complex tasks.
⭐ Stars: 1609
💻 Language: Swift
🔗 Repository: steipete/Peekaboo