Skip to content

Voice and Hearing

Voice and hearing settings let the Virts listen to voice notes, transcribe them instantly, and respond with natural-sounding speech.

Preconditions

  • Editor permissions in Worken
  • A channel that supports audio (phone line, WhatsApp voice notes, etc.)
  • Backup text scenario in case audio delivery fails

Steps

01

Open the “Voice & Hearing” section

Select the Virts, go to the Chats tab, and open Voice & Hearing. Here you can control how the Virts reacts to voice vs. text messages and how it speaks back to the user.

Voice settings pageVoice settings page
Both transcription and speech synthesis live on the same screen
02

Configure speech recognition

Turn on the voice-message reaction, confirm the fallback behavior for text messages, and verify the STT language. This ensures the Virts automatically transcribes every audio snippet without routing it to a live agent.

Speech recognition togglesSpeech recognition toggles
Enable reactions for both voice and text if the channel supports them
03

Pick a TTS model and voice

Choose the voice timbre that matches your brand and set the TTS model. Save the settings and send a test phrase to hear the result directly in the console.

Voice selection blockVoice selection block
tts-1 is ideal for tests, tts-1-hd for production phone calls

WARNING

Voice replies consume more tokens and Workens than plain text. Keep an eye on daily budgets or set spending alerts before rolling audio assistants into production campaigns.

Practical Example

Example

Practical Example

A retail bank enables voice notes in WhatsApp. Customers dictate the last digits of their card, the Virts transcribes the input, checks the request through integrations, and answers with tts-1-hd so the reply sounds natural on the phone line.

Руководство пользователя Worken AI