Peekaboo

View on GitHub

Peekaboo is a macOS CLI & optional MCP server that enables AI agents to capture screenshots of applications, or the entire system, with optional visual question answering through local or remote AI models.

★ 5.2K405 forksSwiftMITopenclaw

Peekaboo is a macOS CLI and optional MCP server that lets AI agents capture screenshots, inspect app accessibility trees, and click or type in native UIs. It supports local or remote model-based visual Q&A and exposes its tools to MCP clients like Claude Code or Cursor.

Use Cases

Capture screenshots of apps or the full system for AI agentsInspect macOS accessibility trees to identify UI elementsAutomate native macOS UI actions like click, type, scroll, dragAnswer visual questions about screenshots using local or remote modelsExpose macOS automation tools to MCP clients such as Claude Code, Cursor, or CodexRun natural-language multi-step desktop automation agentsDebug and visually verify macOS application workflowsControl app windows, menus, dialogs, dock, and spaces from CLIBuild computer-use agents that observe and interact with macOSTake targeted window screenshots for visual inspectionIntegrate screen capture and UI automation into AI coding workflowsProvide screenshot and accessibility context to LLM-based agents

Built With

Language: Swift
Frameworks: Swift · SwiftUI · AppKit · Node.js · MCP · Chrome DevTools MCP

Peekaboo

Use Cases

Built With

Tags