Peekaboo

Peekaboo is a powerful macOS Command Line Interface (CLI) and optional Model Context Protocol (MCP) server designed for developers who want to enhance their AI applications with visual capabilities. This tool allows you to capture screenshots of specific applications or your entire system, and it supports visual question answering using local or remote AI models. By integrating Peekaboo into your workflow, you can streamline the process of collecting and analyzing visual data for your AI solutions.

Once you have connected to the Peekaboo server, you can interact with it using various commands to capture screenshots or initiate visual question answering. For optimal results, use specific queries related to the application you want to capture, such as screenshot <app_name> for capturing an application screenshot or vqa <question> for visual question answering. Ensure your AI models are correctly set up to handle these requests effectively.

To install Peekaboo, ensure you have the following prerequisites:

• macOS
• Node.js (latest version recommended)
• Python (for AI model integration)

Option A: Quick Start with npx

If you want to quickly try out Peekaboo, you can use npx:

npx -y steipete/Peekaboo  

Option B: Global Install Alternative

For a global installation, you can clone the repository and install it directly:

git clone https://github.com/steipete/Peekaboo.git  
cd Peekaboo  
npm install -g  

When configuring Peekaboo, ensure you set the correct environment variables for your AI models, as this can affect your ability to perform visual question answering effectively. Common environment variables include MODEL_PATH for your AI model location. A common gotcha is forgetting to grant screen recording permissions for Peekaboo in your macOS settings, which is necessary for screenshot functionality to work.