Voice Agent

const session = await RunAnywhere.voice.createSession({
  stt: { id: 'whisper-base' },
  llm: { id: 'qwen3-0.6b' },
  tts: { id: 'piper-en-amy' },
})

await session.start()

interface VoiceSessionOptions {
  stt: ModelRef;
  llm: ModelRef;
  tts: ModelRef;
  vad?: VadOptions;
  turnHandling?: TurnHandlingOptions;
  generation?: LlmOptions;
  downloadIfNeeded?: boolean;   // default true
}

createSession(options: VoiceSessionOptions): Promise<VoiceSession>

session.start(): Promise<void>
session.say(text: string): Promise<SpeechHandle>
session.interrupt(): Promise<void>
session.close(): Promise<void>

const session = await RunAnywhere.voice.createSession({
  stt: { id: 'whisper-base' },
  llm: { id: 'qwen3-0.6b' },
  tts: { id: 'piper-en-amy' },
  turnHandling: {
    endpointing: { minDelayMs: 500, maxDelayMs: 3000 },
    interruption: { enabled: true, minDurationMs: 500 },
  },
})

generation: {
  systemPrompt: 'You are a voice assistant. Answer in one or two sentences.',
  maxOutputTokens: 100,
}

import { RunAnywhere } from '@runanywhere/web'

let session: Awaited<ReturnType<typeof RunAnywhere.voice.createSession>> | null = null

startButton.addEventListener('click', async () => {
  session = await RunAnywhere.voice.createSession({
    stt: { id: 'whisper-base' },
    llm: { id: 'qwen3-0.6b' },
    tts: { id: 'piper-en-amy' },
    generation: {
      systemPrompt: 'Answer in one or two sentences.',
      maxOutputTokens: 100,
    },
  })

  await session.start()
  await session.say('Hello. What can I help with?')
})

stopButton.addEventListener('click', async () => {
  await session?.close()
  session = null
})

window.addEventListener('pagehide', () => {
  void session?.close()
})

const result = await RunAnywhere.diarization.diarize(audio)