Open Source
Explore the latest AI open-source projects from GitHub and HuggingFace.
Explore the latest AI open-source projects from GitHub and HuggingFace.

Mobile-Agent is a powerful GUI agent family developed by Alibaba's Tongyi Lab (X-PLUG) that enables autonomous cross-platform device operation through visual perception and language understanding. The project supports Android mobile devices, Windows/Linux desktops, and web browsers, making it one of the most comprehensive GUI automation frameworks available. The latest release, GUI-Owl 1.5 (February 2026), introduces native multi-platform foundation models ranging from 2B to 235B parameters that achieve state-of-the-art results on over 20 GUI benchmarks. The repository contains multiple interconnected agent versions including Mobile-Agent-v3.5 (current flagship), UI-S1 (reinforcement learning variant), GUI-Critic-R1 (error diagnosis), and PC-Agent (desktop-focused). Key architectural features include advanced planning systems, memory management, reflection mechanisms, and MCP (Model Context Protocol) integration for external tool invocation. The project has earned recognition at NeurIPS 2024 and received best demo awards at CCL 2024 and 2025. With 7,500+ GitHub stars and active development, Mobile-Agent represents a significant advancement in autonomous GUI interaction and device automation.