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

# Agent Voice

> Voice-to-Text Communication for Agents

## Agent Voice

Agent Voice to text enables users to communicate with agents using speech input, creating more natural and intuitive conversational experiences. This feature transcribes spoken input into text in real-time, allowing agents to process and respond to voice messages just like typed input.

Agent Voice represents a key step toward creating conversational AI experiences that feel as natural as talking to a human colleague, removing friction from real-world conversations.

## Demo

<video controls className="w-full aspect-video rounded-xl" src="https://storage.googleapis.com/docs-mindset-ai-videos/agent-voice-demo.mp4" />

## Key Features

### Voice Input Functionality

* Users can click the microphone button to start speaking to their agents
* Speech is transcribed in real-time with visual feedback
* Transcribed speech is converted to text messages that agents process like typed input

### Real-time Transcription

* Words appear instantly as users speak
* Powered by Assembly AI for accurate speech recognition
* Visual indicators show when voice recording is active

### Conversation Control

* Users maintain complete control over when messages are sent
* Transcribed speech can be reviewed and edited before sending
* Users must press send to confirm their message after speaking
* Users can interact with the Agent continuously without needing to press the mic button repeatedly, leading to a more natural conversational experience
* Seamless switching between voice input and typing within the same conversation

### Visual Feedback System

* Clear visual indicators show when voice recording is active
* Animated microphone button provides intuitive user feedback
* Real-time transcription display keeps users informed of what's being captured

## Use Cases for Your End Users

* Let your users send rambling voice notes to your AI so there is no friction between an idea, a question, or a challenge, and a helpful answer
* Your users can have their ideas appear in text format, in real time
* Your users can ‘chat’ with AI constantly without needing to click buttons

## How to Monetize This Feature

* At the user level, you can control who has access based on certain packages they purchase, upsell, and more.

## Current Limitations

### Language Support

* Agent Voice currently supports the English language only

### Voice Response

* This is a voice-to-text solution only

### SDK Requirements

* Exclusively available to customers using SDK 2
* Requires migration to SDK 2.0 if not already completed

## Getting Started

### Prerequisites

* Must be using SDK 2.0 or later

### Enabling Agent Voice

* To enable Agent Voice, please check the [mindset.init() method parameters](/sdk-api/sdk/sdk-tags-methods-details#mindset-init-method) page for more details