> ## Documentation Index
> Fetch the complete documentation index at: https://ago.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Voice Conversations

> Talk to the assistant in a live voice call — the conversation appears in the chat as you speak and stays in your history

Voice Conversations let users talk to the assistant instead of typing. Click **Talk** in the message composer to start a live call: you speak, the assistant answers out loud, and the whole exchange appears in the chat thread in real time — just like a typed conversation.

Voice is available in the client portal when it is enabled for your organization.

## How It Works

1. Click **Talk** in the message composer. The first time, a consent dialog explains that your voice is streamed to the AI provider and transcribed into the chat.
2. A compact voice bar appears above the composer showing the call status ("Listening…", "One moment…", "Speaking…") and a microphone level meter so you know you are being heard.
3. As you talk, your words and the assistant's replies appear live in the thread. Your turn shows up once you finish speaking; the assistant's reply streams in as it speaks.
4. Use the microphone button to mute or unmute yourself, and the hang-up button (or the Escape key) to end the call. While the call is live, the text input is disabled.
5. When the call ends, the transcript stays in the conversation — you can keep typing in the same thread, share it, or find it later in your history.

## The Conversation Record

* **Saved like any conversation** — every spoken turn is stored in the thread and visible in the conversation history, including staff and analytics views. Call boundaries are marked with "Voice call started" and "Voice call ended" events showing the call duration and number of turns.
* **The full answer, not just the spoken summary** — when the voice assistant consults the main assistant for a substantive answer, that consultation appears under the spoken reply as an expandable card showing the question asked and the complete written answer.
* **Transcription marker** — voice turns carry a small microphone badge. Hovering it shows a reminder that the text was transcribed automatically from speech and may contain errors.
* **Interruptions are honest** — if you talk over the assistant, its reply stops and the transcript is marked **Interrupted**, so the record reflects what was actually said, not what the assistant would have said.
* **Save failures are visible** — in the rare case a turn could not be saved, it is flagged inline ("Not saved") and the call-ended event shows how many turns were affected. Nothing disappears silently.

## Notes

* Voice requires a secure (HTTPS) connection and microphone permission in the browser.
* Interrupting the assistant mid-sentence is supported — just start talking.
* Phone-based voice (calls to a phone number) keeps its previous behavior; this page describes in-chat voice conversations.