VoiceMem tested by Argusic Runner
Attempt 2 of 3 · Runs with mocks · Argusic Score 92.0 of 100
all runs of VoiceMemaiai-agentsai-toolsapplicationaudio-streaminggenai
Session recording
Download the original recording · fingerprint 9a0d0e8162043bea...
What the test found
VoiceMem package installs and builds in a venv, 5 unit tests pass, and the left-brain ingest+search cycle works end-to-end against a mock OpenAI-compatible server returning the correct schema.
VoiceMem works end-to-end with a schema-correct mock OpenAI API server. Left-brain ingest+search cycle verified. Right brain requires runtime user reactions for trait learning. VAD component skipped during warmup due to missing model file (sherpa-onnx model not downloaded). ASR (FunASR paraformer) downloads and loads successfully. Local E5 embedding (384-dim) works. All 5 existing unit tests pass.
Argusic installed VoiceMem in 13 minutes and launched it; verification reached Verified with mock services. 3 of 3 recorded errors were worked through (see the timeline below).
Runs with mocksall runs of VoiceMem
This summary is drawn from the agent's recorded report for this run. Every figure traces to the log and recording above; nothing here is authored.
Error and fix timeline
No errors were recorded for this run.