google

ARTEMIS: Automate Android Phones with AI Assistants

ARTEMIS lets AI assistants and test suites automate real Android phones from natural-language instructions. Its MCP server, CLI, web console, and Python SDK support device workflows, testing, and diagnostics.

NewPythonApache-2.0AndroidTestingMCP
View on GitHub Compare
Stars
17.3k
Forks
1.9k
Last commit
1 month ago
Contributors
6
Releases, 12 months
0

Our take

Newless than six months old and changing fast

Health score

52/100

How it is scored
Activity
53
Community
25
Issues
67
Pull requests
62

ARTEMIS is a fast-moving Android automation project with broad interfaces for testing and AI-assisted device control, but its young age and highly concentrated contributions make adoption a project-level risk to assess.

Good fit if

  • Your work targets Android phones or emulators and benefits from natural-language automation or MCP integration.
  • You can run Python 3.12+ and provide a device with USB debugging, and you are comfortable with a project that has no recent release history.
  • Apache-2.0 licensing fits your use, and you can account for the project’s reliance on a small contributor group, with most contributions concentrated in one person.

Look elsewhere if

  • You need iOS automation today, which the README lists as a roadmap item rather than a supported platform.
  • Your team requires a mature release cadence or broadly distributed project ownership before adopting a test dependency.
  • Your workflows require the Flash profile’s planning, checkpoint verification, or ADB shell access, which the README identifies as Flash limitations.
All health signals
Last commit2026-09-29 (1 month ago)
Commits, last 90 days100+
Releases, last 12 months0
Contributors6 (top contributor: 92% of commits)
Issues closed, last 90 days12 (typically closed in 2 days)
Pull requests merged, last 90 days6 (typically merged in 0 days)
Project age3 months

Checked on 2026-11-12 with the GitHub API.

Overview

ARTEMIS is an Android automation tool for AI assistants and test suites. It turns natural-language instructions into actions on real phones or emulators, helping teams run end-to-end workflows and gather diagnostics without scripting every UI interaction. It connects to AI coding assistants through a native Model Context Protocol (MCP) server, and also offers a CLI, web console, and Python SDK.

The project combines Android accessibility information, OCR, and visual targeting to interact with different kinds of interfaces. Its Flash profile prioritizes quick reactive actions, while Pro adds planning, action checks, and verification for longer workflows. The README reports 99%+ completion on Google Research's AndroidWorld benchmark; that is a project-reported result, not an independent guarantee for other apps or devices.

Key Features

  • Run cross-app Android workflows from natural-language prompts on a connected phone or emulator.
  • Connect MCP-compatible AI assistants to devices for task execution, screenshots, and Logcat diagnostics.
  • Choose Flash for faster reactive tasks or Pro for planning, action checks, and verification.
  • Locate controls using accessibility hierarchies and OCR, with visual and coordinate fallbacks for custom interfaces.
  • Use the web console for live device viewing, task dispatch, execution telemetry, and replay.
  • Integrate automation into Python test suites or CI/CD pipelines with the Python SDK and structured results.

Use Cases

  • Mobile QA engineers can ask an AI coding assistant to reproduce a bug on a connected device and return screenshots and logs.
  • Test teams can automate multi-step flows across Android apps when they need natural-language dispatch and execution traces.
  • Developers can use the CLI or Python SDK to run device checks in local workflows or CI pipelines.
  • Teams exploring app behavior or stability can use Pro's longer-running workflows, checkpoints, and optional reports.

What you need

Detected in the repository

  • Python >=3.12 (from pyproject.toml)
  • A Dockerfile, so it can run in a container
  • A test suite and automated checks on GitHub Actions

License in plain words

Apache-2.0permissive

  • Commercial use: yes
  • Modify and redistribute: yes
  • You must keep: the license, the NOTICE file and a note of your changes
  • Share your changes: no
  • Includes an explicit patent grant from the contributors:

A summary, not legal advice: the LICENSE file is what applies.

Getting Started

Connect an Android device with USB debugging enabled, then run:

git clone https://github.com/google/artemis.git && cd artemis
./start.sh

On Windows, use start.bat. See the README for setup details, MCP configuration, and other usage modes.

Alternatives

  • android-mcp-server: It provides MCP-based ADB control and device inspection, while Artemis also offers a CLI, web console, Python SDK, and broader testing workflows.
ProjectLanguageLicenseStarsStatus
ARTEMISPythonApache-2.017.3kNew
android-mcp-serverPythonApache-2.0811Inactive

Considerations

ARTEMIS is marked as new and is changing quickly, with substantial recent commit activity, but it has not published a release in over a year. Its six contributors are not evenly represented, with most contributions coming from one person; issue handling has been relatively quick, but that does not remove the continuity risk. The project requires Python 3.12 or newer and a connected Android device or emulator. The README also describes iOS support as planned, not available. The project is licensed under Apache-2.0.

Found this useful?

Share it with someone who would like ARTEMIS.