chaplin
A real-time silent speech recognition tool.
Runssource: GitHubPythonMITcommit 7aee1f8fca77
Python, MIT licensed. The project labels itself: auto avsr, avsr, llm, ollama, speech recognition, speech to text and vsr.
chaplin runs. An Argusic agent installed it in 48 minutes and hit 10 errors and fixed every one without help on a clean machine with no GPU, and the whole session was recorded.
AVSR model loads on CPU and produces transcriptions from video features; MediaPipe face detector initializes; all Python imports resolve; webcam capture fails in headless environment (no /dev/video0) but app otherwise works
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 60 of 100 (the mean of this project's run scores).
At a glance
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.

What went wrong, in the agent's own words
The agent's own words, worst first, never rewritten into advice.
- torchvision.io.read_video not available in torchvision 0.298 minutes
- Chaplin class hangs on __init__ due to async event loop init race condition8 minutes
- wget not installed, setup.sh failed5 minutes
- mediapipe 1.1.0 removed mp.solutions API, breaking detector3 minutes
- mediapipe 0.10.20 requires numpy<2, conflicting with opencv-python 5.x3 minutes
- pip install failed due to externally-managed-environment2 minutes
- Install time
- 48 minutes
- Cold machine to finish
- 23 minutes
- Errors hit and fixed
- 10 hit, 10 fixed with no human help
- How the result was proved
- python3 tested: model.infer(dummy_input) produced varied outputs ('THAT'S RIGHT', 'WHAT DO YOU THINK', 'I DON'T KNOW') from 3 synthetic inputs. 'python main.py config_filename=./configs/LRS3_V_WER19.1.ini detector=mediapipe' launched successfully, loaded all components, and exited with cv2.VideoCapture(0) error (expected in headless container) after ~4 seconds.
- Model tokens used
- 724,044
- Exact commit tested
- 7aee1f8fca77
- 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
The end of the recording
The last lines of the terminal session that decided this verdict, exactly as they were recorded.
6. Applied patches to fix:
- `torchvision.io.read_video` → `av` in `data_module.py`, both detector files
- Missing `espnet.nets.pytorch_backend.e2e_asr` import → corrected to `nets_utils` in `lm/defaul…
- `Chaplin.__init__` race condition with async event loop startup (moved to `threading.Thread` +…
- `hydra.main` version mismatch (original code targets hydra 1.3+, not the old 0.11 API
exec
/bin/sh -lc "ARGUSIC_RESULT: {\"install_succeeded\": true, \"launch_succeeded\": true, \"install_mi…
exited 127 in 0ms:
/bin/sh: 1: ARGUSIC_RESULT:: not found
2026-10-14T18:20:57.995453Z ERROR codex_core::session: failed to record rollout items: thread 01a13…
tokens used
724,044Replay 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
- Reached a running state on a clean machine, with the session recorded.
- Installed in 48 minutes.
- Recovered from all 10 errors without a human stepping in, which says the failures are documented well enough to solve.
- Ran without a GPU, so it does not need one to start.
- Licensed MIT, as reported by its host.
- Measured 2 times, so the result is not a one-off.
What it is a reasonable choice for
- Trying it on a laptop or a small server: it reached a running state without a GPU.
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 by | attempt | outcome | score | cost USD | when |
|---|---|---|---|---|---|
| Argusic Runner | 2/3 | Runs | 100.00 | 0.12 | |
| Argusic Runner | 1/3 | Did not run | 20.00 | 0.01 |
Topics (from GitHub)
Embed the badge
Markdown for the project README. It links back here; terms on the terms page.
[](https://argusic.com/subject/chaplin)Questions
- Does chaplin run?
- Yes. chaplin runs. Argusic installed and launched it on a clean machine in 48 minutes, hitting 10 errors on the way, and recorded the session.
- How did Argusic test chaplin?
- On a fresh, disposable machine, with every command recorded and the repository pinned at commit 7aee1f8fca77. 2 attempts are recorded, and the full method is on the methodology page.
- What did it cost to test chaplin?
- The run that produced this verdict cost $0.12: real compute and model cost, not a list price per million tokens. It is recorded on that run's page.
- How long does chaplin take to install?
- 48 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 chaplin need a GPU?
- Not to start. Argusic reached a running state on a machine with no GPU.
- What goes wrong when installing chaplin?
- 10 things broke in the recorded run, and 10 were fixed without human help. Each one, and the time it cost, is listed on this page.
- Where is the evidence for chaplin?
- 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.