01Full-Duplex Voice AI
NOVASPEAK
For neurodivergent children.
<500ms
end-to-end conversational latency
5×
automated engagement
TWILIOWEBSOCKETSSTT/TTSAWSREAL-TIME
Novaspeak / runtimeListening
LATENCY
438ms
VAD
endpointed
BARGE-IN
armed
MEMORY
adaptive
LISTEN
UNDERSTAND
RETRIEVE
REASON
VALIDATE
SPEAK
Streaming STT/TTSVAD endpointingBarge-in detectionInterruption cancellation
