Artemis (Google)
By Google. Category: Mobile testing with AI. Homepage
Google's README describes Artemis as an open-source agent that executes testing workflows and everyday tasks on Android from natural-language instructions, on real devices and emulators. A Flash profile uses a reactive observe-and-act loop, and a Pro profile adds planning and verification for long-running exploratory and stability tests. Artemis is used through a web console, a CLI, a Python SDK for pytest and a native MCP server with install helpers for Claude Code, Codex, Cursor and other agents. It runs locally with user-supplied Gemini, Claude, GPT-4o or Qwen-VL API keys. It is Apache-2.0 licensed. iOS support is on the roadmap.
Description written from the vendor's documentation.
Status
Status: Active
Facts
| Fact | Value | Evidence | Source | Verified on |
|---|---|---|---|---|
| Status | Active | Documented | github.com | |
| AI capabilities | Documented | github.com | ||
| Test creation | Documented | github.com | ||
| Test output | Not documented | Not documented | None found | |
| Surfaces | Documented | github.com | ||
| Hosting | Documented | github.com | ||
| Pricing model | Documented | github.com | ||
| Free tier | Yes | Documented | github.com | |
| CI integrations | Not documented | Not documented | None found | |
| Self-healing | None documented | Not documented | None found | |
| Agent integration | Native MCP server (python -m mcp_server) exposing mobile_run_task, mobile_manage_task, mobile_get_device_state, mobile_inspect_trace and mobile_diagnose; install helpers for Antigravity, Claude Code, Codex, Cursor, Windsurf, VS Code, Cline/Roo and OpenClaw | Documented | github.com | |
| Open source | Yes, licence Apache-2.0 | Documented | github.com |
Published prices
No prices are published.
Vendor-stated, verbatim
What Google says about its product, quoted as published and not confirmed in documentation.
“AndroidWorld Results: 99%+ task completion on Google Research's AndroidWorld benchmark (100+ multi-step tasks).”
“Flash Execution: A reactive observe-and-act loop with asynchronous history summaries, typically 3–5s per step.”
“Pro Profile (--profile pro): A planning and verification workflow (~15–40s per step), built as a multi-agent graph.”
Alternatives
Tools in the same category (Mobile testing with AI), ordered by the number of shared creation and surface values, then by name. Each line lists documented facts that differ from Artemis (Google).
- Arbigent: documents output Proprietary runner
- AutoMobile: documents output Proprietary runner; creation by prompt to test
- Autosana: documents output Proprietary runner; hosting cloud
- Digital.ai Testing: documents output Mixed; creation by recording, prompt to test; hosting cloud; status acquired
- Drizz: documents output Proprietary runner; creation by prompt to test; hosting cloud
- FinalRun: documents output Proprietary runner; creation by prompt to test; hosting cloud
Questions
Is this page a review of Artemis (Google)?
No. This page is a fact sheet. Each fact shows its evidence level, its source and the date it was checked. We do not rate or rank tools.
Does Artemis (Google) have a free tier?
Yes. Artemis (Google) has a free tier. This is documented, checked on 2026-09-24.
What test output does Artemis (Google) produce?
The test output format is not documented as of 2026-09-24.
Is Artemis (Google) still available?
Status: Active, checked on 2026-09-24.
Sources
Spotted an outdated or wrong fact about Artemis (Google)? Send the source and we apply it at the next monthly update.