Show HN: AutoCua, one prompt drives computer, browser, phones, code at same time

Category: ai-ml

Tags: computer-use, ai-agents, automation, browser-automation, mobile-automation, accessibility-tree, python

Score: 7.3/10 (Innovation: 7, Technical: 8, Documentation: 7, Utility: 7)

AutoCua is a multi-surface computer-use agent framework that lets a single prompt drive desktop apps, browsers, shell terminals, and iOS/Android devices through real OS-level input events and accessibility-tree fusion rather than pixel guessing. It ships prebuilt Rust agents, a Chrome extension, and a desktop app, with a unified Python API across five execution modes. The project is notable for its attention to avoiding debugging ports and VM-level hacks while still handling long-running tasks via transcript compression.

Target audience: AI engineers, automation developers, RPA practitioners, and power users building agentic workflows across desktop and mobile

Repository: https://github.com/ComputerUseAgent/AutoCua · Python · MIT · 1218 stars

View on Hacker News