google/artemis

ARTEMIS turns natural-language instructions into reliable Android automation. It automates end-to-end workflows, captures logs, and integrates seamlessly with AI coding assistants such as Antigravity, Codex, and Claude Code. It also achieves 99%+ success rate on AndroidWorld Benchmark.

What it solves

Artemis allows AI assistants and automated test suites to interact with real Android devices as a human would. It bridges the gap between natural language instructions and complex mobile UI interactions, enabling autonomous cross-app automation, bug reproduction, and stability testing without requiring manual scripting for every UI element.

How it works

Artemis uses a multimodal "observe-and-act" loop. It combines accessibility hierarchies, OCR, and visual models to locate elements on the screen, falling back to coordinates when necessary. It offers two execution modes:

  • Flash Profile: A fast, reactive loop (3–5s per step) for routine tasks, using history compression to maintain context.
  • Pro Profile: A sophisticated multi-agent graph featuring a Planner (for milestones), an Operator (for execution), and a Checker (for verification). It includes a "Safety Net" that checks targets before actions and handles execution incidents for autonomous recovery.

The system can be integrated via a Web Console, a Python SDK, or as a Model Context Protocol (MCP) server that connects AI IDEs (like Claude Code or Windsurf) directly to a mobile device.

Who it’s for

  • Mobile QA Engineers: For creating autonomous test suites and stability tests.
  • Android Developers: For reproducing bugs and performing diagnostic audits using AI IDEs.
  • AI Researchers: For benchmarking mobile agent performance (e.g., on AndroidWorld).

Highlights

  • High Success Rate: Achieves 99%+ task completion on the AndroidWorld benchmark.
  • MCP Integration: Native support for Model Context Protocol, allowing AI coding assistants to drive real phones and collect Logcat output.
  • Multimodal Targeting: Uses a hybrid approach of element indices, coordinates, and visual locating for custom interfaces (Canvas, Compose, Flutter).
  • Flexible Deployment: Available as a CLI, Web UI, Python SDK, and MCP server.

Related

  • Project
    ls1intum/Artemis

    Artemis is a self‑hosted, MIT‑licensed platform for university‑level interactive learning. It supports many exercise types, automatic grading, built‑in version control/CI, plagiarism detection and adaptive learning. Optional AI services (Iris, Athena, Hyperion, Atlas agents) provide LLM‑based tutoring, feedback suggestions and authoring assistance. Deployable via a single Docker‑Compose command, it’s used by multiple institutions and released bi‑weekly.

  • Project
    minitap-ai/mobile-use

    An open-source AI agent that automates Android and iOS devices using natural language to perform tasks and scrape data from apps.

  • Project
    Core-Mate/OpenGUI

    OpenGUI is a framework that lets AI agents control real Android phones by interpreting screenshots and issuing UI actions. It provides WorkBuddy and DeepSeek Harness plugins for easy chat‑based use, a full optional backend for advanced task planning, and a small Android client that performs taps, swipes, typing, etc. Ideal for AI‑driven testing, automation, and content creation on mobile devices.

  • Project
    yashab-cyber/opendroid

    OpenDroid is an Android app that runs an autonomous AI agent on the phone. It parses natural‑language requests into multi‑step plans, executes actions across system settings, apps, navigation, payments, and smart‑home devices, and verifies results. The app includes an on‑device model manager for LiteRT‑LM models, a vision engine for screen understanding, a four‑tier personal knowledge graph, voice wake‑word support, and a premium Jetpack Compose UI. It supports 13 LLM providers (Gemini, Claude, OpenAI, Ollama, etc.) with automatic fallback. Built with clean architecture, Dagger‑Hilt, Room, and Kotlin, it requires JDK 21 and Android 15 to compile. Licensed under Apache 2.0.

  • Project
    suyiiyii/AutoGLM-GUI

    An AI-powered Android automation tool that allows users to control mobile devices via natural language, schedule recurring tasks, and manage multiple devices through a GUI.