VibeVoice
Open-Source Frontier Voice AI
Runs with mockssource: GitHubhomepagePythonMITcommit 16fb2cb1217c
Python, MIT licensed.
VibeVoice runs, with stand-ins for the services it depends on. An Argusic agent installed it in 13 minutes and hit 3 errors and fixed every one without help on a clean machine with no GPU, and the whole session was recorded.
The repo installs and builds in a venv, the 86-test pytest suite passes, the vLLM plugin registers successfully, the Gradio web demo launches and serves HTTP 200, and the streaming ASR server's full API/WebSocket transport works end to end with a schema-correct mock engine in place of the GPU checkpoint.
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
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.
- vLLM 0.14.1 wheel missing runtime dependencies (zmq, cbor2, gguf, aiohttp, msgspec, blake3, cachetools, pybase64, cpuinfo, prometheus-client, openai, openai-harmony, mistral-common, llguidance, xgrammar, diskcache, lm-format-enforcer)8 minutes
- Gradio launch produced empty log and exited when started from a non-interactive shell (first attempt)4 minutes
- pip install -e . refused on system Python: externally-managed-environment (PEP 668)2 minutes
- Install time
- 13 minutes
- Cold machine to finish
- 17 minutes
- Errors hit and fixed
- 3 hit, 3 fixed with no human help
- How the result was proved
- python -m pytest tests/ -q exited 0 with 86 passed; real Gradio demo process answered HTTP 200 on http://127.0.0.1:7860; mock-engine FastAPI harness returned 200 on /health, /v1/models, /v1/config, /v1/transcribe, /v1/transcribe_batch, /v1/chat/completions (plain and SSE) and WebSocket /v1/stream with SRT; vllm.general_plugins entry point loaded and register_vibevoice() ran without error
- Model tokens used
- 1,053,878
- Exact commit tested
- 16fb2cb1217c
- 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
The end of the recording
The last lines of the terminal session that decided this verdict, exactly as they were recorded.
- Ran the repo's pytest suite: `86 passed in 1.01s`. - Verified the vLLM plugin entry point: `EntryPoint(name='vibevoice', value='vllm_plugin:register_v… - Launched the real Gradio demo as a process; it came up and answered `HTTP 200` on `localhost:7860… - Exercised the streaming ASR server transport against a schema-correct mock for the `VibeVoiceAsyn… **Errors fixed (each once, then verified)** - `pip install -e .` failed on the system Python with PEP 668 `externally-managed-environment`; fix… - vLLM 0.14.1's bare wheel missed many runtime dependencies (`zmq`, `cbor2`, `gguf`, `aiohttp`, `ms… **Not verified (gaps, honest)** - Real end-to-end speech transcription on a GPU checkpoint, because this container has no GPU (`nvi… **Final state** ``` ```
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 3 errors without a human stepping in, which says the failures are documented well enough to solve.
- Licensed MIT, as reported by its host.
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 13 minutes to install, slower than the median of the 37 comparable projects Argusic has measured.
- Measured once so far. A single attempt shows it can run, not that it runs reliably.
Of the 38 Python projects Argusic has installed and timed, VibeVoice was the 34th fastest to reach a running state, and 20 of 38 reached one at all.
Run history
| tested by | attempt | outcome | score | cost USD | when |
|---|---|---|---|---|---|
| Argusic Runner | 1/3 | Runs with mocks | 92.00 | 0.19 |
Embed the badge
Markdown for the project README. It links back here; terms on the terms page.
[](https://argusic.com/subject/vibevoice)Questions
- Does VibeVoice run?
- VibeVoice runs, with mocks standing in for external services. Argusic installed and launched it on a clean machine in 13 minutes, hitting 3 errors on the way, and recorded the session.
- How did Argusic test VibeVoice?
- On a fresh, disposable machine, with every command recorded and the repository pinned at commit 16fb2cb1217c. One attempt is recorded, and the full method is on the methodology page.
- What did it cost to test VibeVoice?
- The run that produced this verdict cost $0.19: real compute and model cost, not a list price per million tokens. It is recorded on that run's page.
- How long does VibeVoice take to install?
- 13 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 VibeVoice need a GPU?
- Not to start. Argusic reached a running state on a machine with no GPU.
- What goes wrong when installing VibeVoice?
- 3 things broke in the recorded run, and 3 were fixed without human help. Each one, and the time it cost, is listed on this page.
- How does VibeVoice compare with the alternatives?
- Of the 38 Python projects Argusic has installed and timed, VibeVoice was the 34th fastest to reach a running state, and 20 of 38 reached one at all.
- Where is the evidence for VibeVoice?
- The recorded run is on this page, each linking to its full log and terminal recording, stored with a sha256 fingerprint so it cannot be quietly altered.