v2026.7.15
Release date: July 27, 2026
New Features
Automatic audio transcription (per-channel settings)
Conversation audio can be automatically converted to text with AI. The transcript appears on the message and is also used as context by the AI Agent and assistant.
Channels now have separate controls:
| Setting | Default | What it does |
|---|---|---|
| Transcribe customer audio | On | Transcribes audio sent by the customer |
| Transcribe agent audio | Off | Transcribes audio sent by the agent (Interflow or channel app) — opt-in |
| Transcribe group audio (WhatsApp WAHA only) | Off | Allows transcription in group chats as well |
What is it for?
- Read what was said without listening to every audio
- Speed up replies and text search in history
- Give the AI text context even when the customer sends voice notes
- Control cost and usage: customer on by default; agent and groups only when needed
How to use?
- Open Channels and edit the desired channel
- Enable or disable Transcribe customer audio and Transcribe agent audio
- On WhatsApp WAHA, if groups are enabled, use Transcribe group audio when it makes sense
- Ensure an active OpenAI integration or AI credits balance
- For audio sent from the agent's phone, also keep external media upload enabled on the channel
Existing channels
With no saved setting: customer audio keeps being transcribed; agent and group audio stay off until you enable them.
Full guide: Channels — Audio transcription
Benefits
- Less time spent replaying voice messages
- Searchable text history for agents and AI
- Per-channel controls to balance quality and credit cost
- WAHA groups without unnecessary transcription by default