Sayd

VAD API — Voice Activity Detection

Detect whether audio contains speech and find exact speech segments. Useful for pre-filtering audio before transcription, or trimming silence from recordings.

Python
from sayd_ai import Sayd

client = Sayd(api_key="sk-your-key")

# Detect speech segments in audio
segments = client.vad.detect("recording.wav")
for seg in segments:
    print(f"Speech: {seg['start']:.2f}s - {seg['end']:.2f}s")

# Quick check: does this audio contain speech?
has_speech = client.vad.check("recording.wav")
print(f"Has speech: {has_speech}")  # True / False

# Also accepts raw bytes
with open("audio.pcm", "rb") as f:
    has_speech = client.vad.check(f.read())

API Endpoints

POST/v1/vadDetect speech segments in audio
POST/v1/vad/checkCheck if audio contains speech