voice_assistant

Hailo Voice Assistant Interactive Application

An interactive voice-controlled AI assistant using Hailo's Speech-to-Text and Large Language Model for real-time audio processing and conversational AI.

Prerequisites

Required: Piper TTS Model Installation

Before running the voice assistant, you must install the Piper TTS voice model.

For installation instructions, see:

• Voice Processing Module Documentation

The application will check for the model on startup and display an error with instructions if it's missing.

The voice assistant uses the shared voice processing module from hailo_apps.python.gen_ai_apps.gen_ai_utils.voice_processing.

Audio Setup

High-quality microphone input is crucial for optimal speech recognition performance.

For audio device configuration and troubleshooting, use the built-in tool:

hailo-audio-troubleshoot

This tool will:

• List all available audio devices
• Test your microphone and speakers
• Help select and save preferred devices
• Provide platform-specific setup instructions (Raspberry Pi, etc.)

For detailed audio setup instructions, see: Voice Processing Module Documentation

Tips for best results:

• Use a USB microphone for better quality than built-in mics
• Position microphone 15-30cm from your mouth
• Minimize background noise during recording
• Speak clearly at a normal volume

Features

• Real-time speech processing with Hailo AI acceleration
• Interactive voice mode - press Space to start/stop recording
• Streaming text-to-speech - responsive audio playback with interruption support
• Context management - maintains conversation history with clear option
• Debug logging - detailed logging for troubleshooting (use --debug or --log-level DEBUG)
• Low-resource mode - optional TTS disable for reduced system load

Requirements

• Hailo AI processor and SDK
• Python 3.10+
• sounddevice (for audio I/O)
• NumPy
• Piper TTS (for voice synthesis)
• Hailo Platform libraries

Installation

Linux

Before running this example, ensure GenAI dependencies are installed:

# From the repository root directory
pip install -e ".[gen-ai]"

Windows

Use a clean virtual environment before installing the dependencies:

py -m venv wind_venv
.\wind_venv\Scripts\Activate.ps1
pip install .\hailort-<version>-cp<python>-cp<python>-win_amd64.whl

# From the repository root directory
pip install -e ".[gen-ai]"

Files

• voice_assistant.py - Main application with terminal-based voice interface

Usage

1. Run the application:

   python -m hailo_apps.python.gen_ai_apps.voice_assistant.voice_assistant

2. Optional flags:

   python -m hailo_apps.python.gen_ai_apps.voice_assistant.voice_assistant --debug      # Enable debug logging
   python -m hailo_apps.python.gen_ai_apps.voice_assistant.voice_assistant --no-tts     # Disable text-to-speech
   python -m hailo_apps.python.gen_ai_apps.voice_assistant.voice_assistant --log-level DEBUG  # Set log level

3. Interactive controls:

   • Press Space to start/stop recording
   • Press Q to quit the application
   • Press C to clear conversation context
   • Press X to abort generation and speech
   • Speak naturally during recording

How it works

The application uses a threaded architecture to handle:

• Real-time audio capture and processing via VoiceInteractionManager
• Hailo Speech-to-Text transcription
• Large Language Model inference for responses
• Streaming text-to-speech synthesis with interruption support
• Non-blocking user input handling

The voice assistant can engage in natural conversations, answer questions, and provide assistance while maintaining context