Difficulty: ⭐⭐⭐☆☆ | Duration: 45 minutes | Gain: Mastering the complete setup of an embedded AI voice assistant

Today, we're diving into something hardcore—using a Raspberry Pi to create a voice-interactive AI assistant.

Press a button, ask your question, and the AI's response will stream live on the LCD screen. If you want it to speak, you can also enable TTS voice broadcasting.

This is pizero-openclaw, an open-source voice AI assistant running on a Raspberry Pi. We'll walk through getting it up and running step by step.

---

Target Audience

• Developers with 1-5 years of experience interested in embedded systems and AI
• Hobbyists looking to build AI hardware
• Students with a basic understanding of Linux command line

---

Core Dependencies and Environment

Before starting, make sure you have the following:

---

Project Structure

pizero-openclaw/
├── main.py                 # Entry file, main loop
├── display.py              # LCD rendering (status, replies, clock)
├── openclaw_client.py      # OpenClaw gateway HTTP client
├── transcribe_openai.py    # Speech-to-text (OpenAI)
├── tts_openai.py           # Text-to-speech (OpenAI)
├── record_audio.py         # Audio recording (ALSA)
├── button_ptt.py           # Button state machine (press to record, release to send)
├── config.py               # Configuration management (.env loading)
├── sync.sh                 # Deployment script (rsync + systemd)
├── .env.example            # Configuration example
└── requirements.txt        # Python dependencies

---

Hardware Preparation

Let's gather all the necessary hardware.

Essential Hardware

Optional Hardware

• PiSugar Battery — Lithium battery module for use without power
• Enclosure — 3D printed or acrylic cut

---

System Flashing and Initialization

Flashing Raspberry Pi OS

1. Download Raspberry Pi Imager
2. Choose Raspberry Pi OS (Bookworm) 32-bit or 64-bit
3. In advanced settings, enable:
   • SSH
   • Configure Wi-Fi SSID and password
   • Set username pi and password

Once the flashing is complete, insert the TF card into the Raspberry Pi and power it on.

First Connection

# Find Raspberry Pi IP on your computer
arp -a | grep raspberry

# SSH connection (default password is raspberry)
ssh [email protected]

After your first login, it's recommended to update the system:

sudo raspi-config

Configure the following options:

• Extend the file system to the entire TF card
• Set the timezone (Localization → Timezone)
• Enable the SPI interface (Interface Options → SPI)

---

Install Dependencies

Confirm Python Environment

# Check Python version
python3 --version
# Output should be Python 3.11+ 

# If the version is too low, upgrade it
sudo apt update
sudo apt install python3.11 python3.11-venv

Install System Dependencies

sudo apt install python3-numpy python3-pil python3-pip

Install Python Packages

# Clone the project
git clone https://github.com/your-repo/pizero-openclaw.git
cd pizero-openclaw

# Install dependencies
pip install -r requirements.txt

requirements.txt should look something like this:

requests
python-dotenv
numpy
Pillow

---

Install WhisPlay Drivers

This step is crucial as the LCD and buttons require drivers to function.

Install the drivers according to the official PiSugar documentation:

# The driver should be installed in /home/pi/Whisplay/Driver/
# For detailed steps, refer to: https://github.com/PiSugar/whisplay-ai-chatbot

After installation, test if the LCD displays correctly:

cd /home/pi/Whisplay/Driver/
python3 test_display.py

If the screen lights up and displays content, the driver installation is successful.

---

Configure Environment Variables

Copy Configuration Template

cd ~/pizero-openclaw
cp .env.example .env
nano .env

Fill in Necessary Configurations

# Required: OpenAI API Key (for speech-to-text and TTS)
export OPENAI_API_KEY="sk-your-openai-api-key"

# Required: OpenClaw Gateway Token
export OPENCLAW_TOKEN="your-openclaw-gateway-token"

# Optional: OpenClaw Gateway Address
export OPENCLAW_BASE_URL="http://192.168.1.100:18789"

Other Common Configurations

---

Set Up OpenClaw Gateway

The voice assistant needs to connect to the OpenClaw gateway for AI interaction.

Run the Gateway on Raspberry Pi (Recommended)

# Install OpenClaw
npm install -g openclaw@latest
openclaw onboard --install-daemon

Run Gateway on Another Machine

If the performance of your other machine running the Raspberry Pi is limited, you can run the gateway on another computer:

# Run on Mac/server
npm install -g openclaw@latest
openclaw onboard --install-daemon

Configure the Gateway

Edit ~/.openclaw/openclaw.json:

{
  "gateway": {
    "bind": "0.0.0.0",
    "port": 18789,
    "auth": {
      "mode": "token",
      "token": "your-token"
    }
  }
}

# Restart the gateway
openclaw gateway restart

# Get Token
openclaw config get gateway.auth.token

Fill the obtained Token into the previously configured .env file.

---

Test Run

Local Start

With everything configured, it's time to test:

cd ~/pizero-openclaw
python3 main.py

Under normal circumstances, you should see:

1. LCD displaying time, date, battery level, WiFi status
2. The screen shows “Ready” waiting for commands

Voice Interaction Test

1. Long press the button on WhisPlay — start recording
2. Release the button — send the recording
3. Wait a few seconds, the screen will display:
   • Recognized text
   • AI's streaming response
4. If TTS is enabled, you'll also hear the voice broadcasting