AutoGLM-GUI is an open-source graphical interface and productivity layer for the AutoGLM phone agent: it lets an AI model operate Android devices from a chat-style web UI, tapping, swiping and completing tasks on real phones and emulators. Released under the Apache-2.0 license, it has about 1,100 stars on GitHub and is actively developed in 2026.
AutoGLM is an open phone-agent model that understands screenshots and plans actions; AutoGLM-GUI turns it into a usable product. Through a browser interface you describe a task in conversation, watch a live scrcpy screen preview of the device, and can even click and swipe directly in the stream when you want to take over. Version 1.5 pushed it from demo to automation hub: cron-style scheduled tasks, Docker deployment for 24/7 operation, saved conversation history, sub-second interruption of running actions, and multi-device management with isolated state per device. Its layered agent design separates a decision model from a vision model for planning and precise execution, it connects devices wirelessly via QR pairing on Android 11+, and it exposes an MCP server so other AI apps can drive it. Desktop builds for Windows, macOS and Linux remove the setup work entirely.
AutoGLM-GUI is free and open source under the Apache-2.0 license, including ready-made desktop downloads. You supply your own OpenAI-compatible LLM API access for the agent models it drives.
AutoGLM-GUI is for developers and power users who want an AI agent that operates Android phones for testing, repetitive app workflows, scheduled check-ins and hands-free device automation.