Skip to content
Hot storyPast story

Inworld releases SAVA voice activity and turn detection model

1 report1 sourceUpdated 4 d ago

What happened

AI overview

Inworld released SAVA, a semantic acoustic voice activity model that merges voice detection and turn detection into a single model, and it is already running in its voice products. The model outputs two probabilities every 40 milliseconds, for speech and end of turn, and has 25.5M parameters; according to Inworld, on a single CPU thread of an Apple M3 Max, processing each 40-millisecond frame takes less than 3 milliseconds. Inworld says SAVA downloads its weights from Hugging Face on first load; it detects any sound it hears without distinguishing between speakers, so dialogue policies still need to handle interruptions and short backchannels like "mm-hm" themselves. The model is intentionally designed to remain sensitive to quiet and distant speech.

AI-written from the coverage · updated 3 d ago

Coverage timeline

Follow the coverage to see the story from different sides.

Nov 18
  1. Inworld AI BlogPick
    Inworld releases SAVA voice activity and turn detection model

    Inworld released SAVA, a semantic acoustic voice activity model that merges voice activity detection and turn detection into a single model and is already running in its voice products. The model outputs two probabilities every 40 ms, for speech and end of turn, has 25.5M parameters, and processes each 40 ms frame in under 3 ms on a single CPU thread of an Apple M3 Max.

Heat trend for this story

Current (comparable)
3
Peak (comparable)
6Nov 19 02:00
24-hour change
-49%

The trend only compares participants observed continuously and completely, so it may cover fewer than the current heat count. Hover over or tap the chart to see each hour's heat; with a keyboard, use the left and right arrow keys.