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

# Voice mode

> Use hands-free or push-to-talk voice conversations in LMSA to speak messages and hear replies without touching the keyboard.

Voice Mode turns LMSA into a hands-free, full-screen conversation: speak your message, listen to the reply, and keep going without touching the keyboard.

> [!NOTE]
>
> Voice Mode is a **Premium** feature. Free users who tap the microphone icon will be offered an upgrade.

## How to start

**1. Open Voice Mode**

Tap the **microphone icon** in the chat input bar.

**2. Start talking**

The Voice Mode screen opens and listening starts automatically. This is **Hands-free** mode, and it's on by default. Just start talking; there's no need to tap anything to begin speaking.

Voice Mode requires **microphone permission**; you'll be prompted to grant it the first time you use it.

## Hands-free vs push-to-talk

* **Hands-free** (default): LMSA listens automatically, detects when you stop talking, and sends your message on its own. It keeps listening again after every reply, so you can have a full back-and-forth without touching the screen.
* **Push-to-Talk**: Turn off **Hands-free** in Voice Mode's **Settings** to switch to press-and-hold instead. Press and hold the mic button to talk, then release to send.

## Interrupt a reply

Don't wait for the AI to finish. Tap the large circle **once** while it's thinking or speaking to interrupt it and jump straight back in with your next message. In Push-to-Talk mode, press and hold the button again to do the same thing.

## Mute button

The small mic-off button below the main circle mutes your microphone. While muted, the main mic button is locked. Tap the mute button again to unmute and resume.

## Voice Mode settings

Tap the **gear icon** in the top-right of Voice Mode to open its own settings. No need to leave Voice Mode:

* **Hands-free**: Turn automatic listen-and-send on or off.
* **LLM Model**: Switch the active AI model mid-conversation.
* **Voice**: Browse or search Text-to-Speech voices by language and country, and change the one used to read replies aloud.

The voice you pick (in Voice Mode's own Settings or under **Settings > Options > TTS Voice**) is used to read every reply aloud.

## Lock screen media controls

If you enable notification permissions, you can control Voice Mode text-to-speech playback from your lock screen and notification shade. This lets you play, pause, or stop voice responses without unlocking your device, and manage playback from quick settings and connected Bluetooth devices.

## Privacy

Your voice is never recorded, stored, or sent to any cloud service; everything happens on your device.

## Exit Voice Mode

The **close icon** (top-left) exits Voice Mode and returns you to the normal chat screen, with the full conversation history intact. Nothing you said is lost.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.