Talk Using Voice

Voice is the planned in-workspace channel for talking to an employee out loud instead of typing. It is not available to enable yet.

TL;DR

Voice will let you speak to an employee and hear its replies back, in real time, inside the workspace. The channel is planned, not live. There is no voice button in chat today.

How It Works

flowchart LR
  A[You speak] --> B[Speech to text]
  B --> C[Employee]
  C --> D[Text to speech]
  D --> E[You hear the reply]

The plan is for your microphone audio to stream to a speech-to-text service, the transcribed text to go through the same execution pipeline as a typed message, and the reply to stream back as speech, sentence by sentence, so you start hearing it before the employee finishes generating the rest.

What It Can Do

Once released, the plan is:

• Start a real-time voice conversation with an employee, hands-free.
• Have the employee use its normal skills, tools, and memory during a voice call, the same as a typed message.
• Switch between voice and text in the same conversation without losing context.
• Keep the full call as a text transcript in chat history. No audio recordings.

None of this works today. Voice is being redesigned and does not appear as a working button anywhere in the product.

How to Set It Up

No setup is available yet. Voice will appear under Settings > Technical > Channels once released.

Tips and Tricks

Until voice ships, type your message instead. Text messages go through the same employee pipeline, skills, tools, and memory, that voice is planned to use.

Good to Know

The card in Settings > Technical > Channels stays a compact Coming soon placeholder until voice delivery ships. Phone Calls is a separate, also-planned channel for calling an employee from a real phone number; see the Phone Calls guide.

Frequently Asked Questions

Q: Can I talk to my employee out loud right now? A: No. Voice is planned but not released. Use text chat today.

Q: Will voice messages be saved as text? A: Once released, yes. The plan is to persist only the transcript, not the audio, in your conversation history.

Q: Will the employee behave differently over voice than over text? A: No. The plan is for voice to reuse the exact same execution pipeline, skills, tools, and memory as text chat.

Q: Is this the same as Phone Calls? A: No. Voice is an in-workspace, browser-based channel. Phone Calls is a separate planned channel for dialing a real phone number. Neither is live yet.

Q: When will voice be available? A: There is no committed release date. It is paused while the voice flow is being redesigned.