TECHNOLOGY PROFILE · TECH-0244
Silero VAD
Voice activity detection models and tooling.
Reference link only — not reviewed this edition. This profile preserves the prepared catalog’s original descriptions. Advantages depend on the workload; comparisons are editorial, not benchmark results.
Where it has an edge
Lightweight speech/non-speech detection avoids transcribing long silent segments.
Best-fit problems
Segmenting microphone streams and preprocessing audio.
Tradeoffs & cautions
Noise, music and atypical speech can cause false detections.
What to test before choosing
Test real microphones/noise/languages; measure word errors, diarization, turn-taking, latency and voice rights.
Runtime & hardware
Local CPU/GPU or realtime infrastructure
Access & licensing
Code/model/voice licenses vary
Execution
Local / self-hosted
Lifecycle
Not independently audited
Comparison basis
Editorial problem-fit judgement; no head-to-head benchmark performed
Connected concepts
Alternatives to investigate
WebRTC VAD · pyannote.audio