OpenGUI

OpenGUI is a mobile GUI agent framework for Android. It reads real screenshots to understand app screens and then taps, swipes, types and navigates to complete multi-step tasks on a connected phone from a natural-language instruction.

Beyond operating apps, it can test features and record issues with reproduction steps and screenshots, read and organize screen content, fill forms and prepare social media drafts, stopping before publishing when told to. In WorkBuddy's workbench you can watch the phone screen, the steps and the report, and take over or stop a task.

The quickest path is the plugin for WorkBuddy (macOS, version 5.5.3 or later; the current release is a public-testing prerelease) or DeepSeek Harness, which needs no backend. A full self-hosted stack with an Android client can take tasks from Feishu, Telegram, Discord or a REST API. OpenGUI accounts use SMS sign-in that currently requires a mainland-China phone number.