Google ARTEMIS: Giving AI Agents Hands to Control Android Devices
It lets AI agents control a real phone like a human.
You hand the agent a task in plain English and it operates the device end to end: sees the screen, finds the elements, taps, swipes, navigates, verifies the outcome, and reports the result.
It connects to your agent over MCP: Claude Code, Codex, Cursor, Windsurf, Antigravity, VS Code, and more.
What Google reports:
→ 99%+ task completion on AndroidWorld
→ 100+ multi-step tasks
→ ~3–5 seconds per step in Flash mode
→ Targeting that fuses accessibility, OCR, and vision
→ Screenshots, execution traces, and diagnostics on every run
→ A Python SDK to drop into your test pipelines
The real shift: your phone turns into an environment an agent can run.
Tell Codex to build the APK, install it, log in with a test account, walk the checkout flow, flag any unexpected popups, and return screenshots of the failures, and it drives the device itself.
QA, mobile testing, debugging, automation, all of it moves forward. Google just gave AI agents hands for Android.
Link: https://lnkd.in/erxwRxyF
100% open source. Apache 2.0.
Google ARTEMIS: Open-Source Android Automation for AI Agents
Google's ARTEMIS is an open-source project that enables AI agents to interact with Android devices through natural-language instructions.
Instead of manually navigating every screen, an agent can observe the device, identify interface elements, perform actions, and execute multi-step workflows across Android applications.
What ARTEMIS offers
→ Natural-language Android automation
→ MCP integration with AI coding assistants
→ Accessibility-based and multimodal UI targeting
→ Screenshots, execution traces, and diagnostics
→ Flash execution mode with reported 3–5 second step times
→ Python SDK for integration into testing workflows
Google's repository reports 99%+ task completion on the AndroidWorld benchmark, covering more than 100 multi-step tasks. This is a project-reported result and should not be interpreted as a universal success rate for every Android application.
Why this matters
A practical use case is AI-assisted mobile testing.
For example, an agent could be instructed to build an APK, install it on a connected Android device, open a test account, navigate through a checkout workflow, and collect evidence of unexpected popups or failures.
The actual reliability of this workflow depends on the app, device configuration, permissions, test data, and agent execution.
ARTEMIS connects device automation with AI coding environments through MCP, making it possible to combine software development and mobile testing workflows.
The project is available under the Apache License 2.0.
🔗 GitHub: https://github.com/google/artemis
#AI #GenerativeAI #AIAgents #Android #MobileTesting #Automation #MCP #SoftwareTesting #OpenSource