Sato

Sato is an open-source desktop application for macOS that combines AI-powered question answering with a customizable pixel art companion. It enables users to take screenshots of any part of their screen, ask questions in plain English about the captured content, and receive streaming responses from AI models such as Claude, GPT, or local models via Ollama. The application is designed to work across any app or window, providing assistance wherever visual information is present on the screen.

The tool features a selection of five animated pixel art pets, including characters like Max, Sky, Lexi, Rover, and Paris, each with distinct personalities and behaviors. These companions live on the user's desktop, offering visual engagement and interactive feedback, such as walking, flying, or reacting to correct answers. Users can switch between companions instantly, and the system supports adding new sprites through its open-source architecture, although this is not yet available through the user interface.

4 and Mini/Nano variants), and Ollama for local or cloud-based models. API keys are managed securely in the macOS Keychain, and users can select different models for different tasks using Context Profiles, which enable the tool to remember project-specific information across sessions. A chat sidebar is available for extended conversations about a screenshot, and Stealth Mode lets users hide the on-screen companion while retaining full AI functionality.

2 or later) and is licensed under the MIT License. Users pay only for API usage with their chosen AI provider unless running Ollama locally, which enables fully offline, cost-free operation. Sato does not operate its own servers and does not access user data; screenshots are sent directly to the selected AI provider, and with local models, data remains entirely on the user's machine. Sato is developed by Amir Valizadeh and its source code is available on GitHub.