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.

Runs with mockssource: GitHubPythonApache-2.0commit 371aa6df5688

Python, Apache-2.0 licensed. The project labels itself: ai agents, android, google, test automation and testing.

artemis runs, with stand-ins for the services it depends on. An Argusic agent installed it in 67 minutes and hit 8 errors and fixed every one without help on a clean machine with no GPU, and the whole session was recorded.

uv sync --dev succeeded; pytest exits 0 with 2185 passed, 10 skipped, 8 deselected; 'artemis ui' served the web console with HTTP 200 on /, /admin, /docs, and /health; --version answered v1.0; doctor and status CLI commands run, and the ADB daemon started from the bundled binary.

What the agent ended up with on a clean machine, in its own words. How this is measured

Measured by Argusic on a fresh machine every time. Every number links to its evidence. Argusic Score 92 of 100 (the mean of this project's run scores).

At a glance

verdict
Runs with mocks
Argusic Score
92 / 100
cost of the verifying run
$1.64 (measured)
recorded runs
2
last tested
stars
17,515
forks
1,951
open issues
191
watchers
231
size
25 MB
created
last push

Subject data from GitHub, linked at the top of this page, refreshed . Test data by Argusic (CC BY 4.0); every number links to a run page with the full log, the recording, and their sha256 hashes.

What was measured

From the run that decided this verdict. Every line is a recorded value, not an estimate.

Time lost to each failure while testing artemis, drawn from the recorded run
Every bar is a failure the agent recorded, with the time it cost. Drawn from this run, not from anywhere else.

What went wrong, in the agent's own words

The agent's own words, worst first, never rewritten into advice.

  • VisualStepSummarizer.__init__ crashed without an API key: except-branch re-called get_google_llm() after the first call failed, and logger calls used printf-style args unsupported by ArtemisLogger15 minutes
  • No real Google API key, Android device, or scrcpy in the container; adb daemon comes from the bundled adbutils binary15 minutes
  • Generated FastMCP tool schemas for ADB server drifted from the pinned fixture: 6 docstrings were indented/missing blank lines and contained a literal backslash-n10 minutes
  • test_unified_controller_video.py relied on hardcoded decoded frame counts/indices that vary by codec/OpenCV build (77 frames decoded for a clip ffmpeg reports as 34-45)8 minutes
  • Unit tests using MagicMock providers hit 'Unknown LLM provider' exception6 minutes
  • subprocess.CREATE_NEW_PROCESS_GROUP/DETACHED_PROCESS are absent on Linux, breaking device utils import and one unit test5 minutes
Install time
67 minutes
Cold machine to finish
61 minutes
Errors hit and fixed
8 hit, 8 fixed with no human help
How the result was proved
uv sync --dev exited 0; timeout 3000 uv run pytest -q exited 0 with '2185 passed, 10 skipped, 8 deselected, 88 warnings in 64.67s'; uv run artemis ui --no-open --port 8123 stayed up and curl returned HTTP 200 on /, /admin, /docs, /health with a compiled Angular index.html on /; uv run artemis --version printed 'Artemis Agent Platform v1.0'; uv run artemis status exited 0 printing OFFLINE/STOPPED board
Model tokens used
7,765,466
Exact commit tested
371aa6df5688
Container image
sha256:cdd920bce783

What was not measured

The other half, so you can see exactly how far this evidence goes. A gap is printed as a gap and never filled in.

Peak GPU memory
this ran on a CPU machine, so there was no GPU to sample
Behaviour over a long run
each attempt uses the smallest documented settings and a fixed time budget, so this measures whether it runs, not how it holds up under load
Behaviour against real services
this run was verified against stand-in services only

Watch the session

The whole run, from a clean machine, at the speed it actually happened. Nothing is cut and nothing is sped up.

A recording of attempt 2, run , the attempt that decided the verdict above. The terminal recording it was made from is the evidence, stored with a sha256 fingerprint; the video is a rendering of it. Nothing loads from the video host until you press play.

The end of the recording

The last lines of the terminal session that decided this verdict, exactly as they were recorded.

All verification is complete. Final state confirmed:
- `uv sync --dev` succeeded; `uv run pytest -q` exits 0 with `2185 passed, 10 skipped, 8 deselected…
- `uv run artemis ui --no-open --port 8123` stayed up ~90s, serving HTTP 200 on `/`, `/admin`, `/do…
- `uv run artemis --version` and `artemis doctor`/`artemis status` all run cleanly.
The one unfixable gap is the Showcase UI auto-rebuild (`ng build` requires Node ≥22.22.3; container…
tokens used
7,765,466
All verification is complete. Final state confirmed:
- `uv sync --dev` succeeded; `uv run pytest -q` exits 0 with `2185 passed, 10 skipped, 8 deselected…
- `uv run artemis ui --no-open --port 8123` stayed up ~90s, serving HTTP 200 on `/`, `/admin`, `/do…
- `uv run artemis --version` and `artemis doctor`/`artemis status` all run cleanly.
The one unfixable gap is the Showcase UI auto-rebuild (`ng build` requires Node ≥22.22.3; container…

Replay the whole session, every command from a clean machine to this point.

How it was tested

One agent, one container, no human help. It installs the project the way its own documentation says to, runs it, and fixes what breaks. Everything below is recorded as it happened: the terminal session, the log and the exact commit. The full method.

Strengths and limits

Measured facts, not opinions. How this is written.

What went well

  • Recovered from all 8 errors without a human stepping in, which says the failures are documented well enough to solve.
  • Licensed Apache-2.0, as reported by its host.
  • Measured 2 times, so the result is not a one-off.

What did not, or is not known

  • Only reached a running state against stand-in services, so nothing here proves it works against the real ones.
  • Took 67 minutes to install, slower than the median of the 121 comparable projects Argusic has measured.

Of the 122 ai-agents projects Argusic has installed and timed, artemis was the 120th fastest to reach a running state, and 94 of 122 reached one at all.

Also tested, in the same area

Every one of these was installed and run by Argusic on a clean machine. Nothing appears here that was not tested.

Run history

tested byattemptoutcomescorecost USDwhen
Argusic Runner2/3Runs with mocks92.001.64
Argusic Runner1/3Runs with mocks92.000.10

Topics (from GitHub)

ai-agentsandroidgoogletest-automationtesting

Embed the badge

Markdown for the project README. It links back here; terms on the terms page.

[![Tested by Argusic](https://argusic.com/badge/artemis.svg)](https://argusic.com/subject/artemis)

Questions

Does artemis run?
artemis runs, with mocks standing in for external services. Argusic installed and launched it on a clean machine in 67 minutes, hitting 8 errors on the way, and recorded the session.
How did Argusic test artemis?
On a fresh, disposable machine, with every command recorded and the repository pinned at commit 371aa6df5688. 2 attempts are recorded, and the full method is on the methodology page.
What did it cost to test artemis?
The run that produced this verdict cost $1.64: real compute and model cost, not a list price per million tokens. It is recorded on that run's page.
How long does artemis take to install?
67 minutes in the run that decided this verdict, from a clean machine to a running state. That is one measured attempt, not an average.
Does artemis need a GPU?
Not to start. Argusic reached a running state on a machine with no GPU.
What goes wrong when installing artemis?
8 things broke in the recorded run, and 8 were fixed without human help. Each one, and the time it cost, is listed on this page.
How does artemis compare with the alternatives?
Of the 122 ai-agents projects Argusic has installed and timed, artemis was the 120th fastest to reach a running state, and 94 of 122 reached one at all.
Where is the evidence for artemis?
All 2 recorded runs are on this page, each linking to its full log and terminal recording, stored with a sha256 fingerprint so it cannot be quietly altered.

Discussion