v2026.8.24
Release date: August 27, 2026
New Features
System voices and audio from text
Save organization voices (OpenAI, ElevenLabs, or Minimax) and, in the conversation, turn the typed text into audio — without recording on the microphone.
The Record from text button sits on the input bar. With one active voice, it generates immediately; with two or more, the attendant picks a voice. The script stays in the field (collapsed) so they can edit or regenerate. On send, the chat stores one message: the audio and the text as the transcription.
Full guide: System voices
What is it for?
- Send audio in the brand voice without recording every message
- Keep tone and speed consistent across attendants
- Keep the script visible in Interflow as a transcription
How to use?
- In Settings, add a TTS integration (OpenAI, ElevenLabs, or Minimax)
- Open Voices, create the voice (name, integration, and settings), and leave it active
- In the chat, type the text and click Record from text
- Preview, edit or regenerate if needed, then send
Who configures it
Admin and owner create, edit, and delete voices. Any member can use active voices in conversations, on channels that accept audio.
Benefits
- Ready audio from text, in the organization voice
- A single message: the customer hears the audio; in Interflow the text is the transcription
- Same providers already used on the flow audio nodes