Availability & Kits (Important)
Bob is free to make and print for personal use. For convenience, a limited number of pre‑printed kits (including screens, switches, and microphones) are available here ominousindustries.com.
- 3D Printed Parts
- Step 1 — Audio, Bluetooth & Chatbot Base
- Step 2 — SPI Display (Waveshare) & “Bob” Chat
- Quick Reference
- Notes
All print files are available here:
Printables: https://www.printables.com/model/1404129-bob-the-sentient-washing-machine
A limited number of pre‑printed kits (including screens, switches, and microphones) are available here ominousindustries.com.
# Update system packages
sudo apt update
sudo apt upgrade -y
# Core tools, audio stack, BT, and TTS
sudo apt install -y \
git python3-venv python3-pip \
python3-gpiozero python3-rpi.gpio \
bluez \
pipewire pipewire-pulse wireplumber \
espeak-ng \
libportaudio2 \
alsa-utils \
libspa-0.2-bluetooth
# Add user to audio group for device permissions
sudo usermod -aG audio $USER
# Reboot for group membership to take effect
sudo rebootAfter reboot, log back in and continue.
# Enable Bluetooth system service
sudo systemctl enable --now bluetooth
# Start PipeWire audio stack for your user
systemctl --user enable --now pipewire wireplumber pipewire-pulse
# Keep user services running even when not logged in (headless)
sudo loginctl enable-linger $USER
# Verify PipeWire is up
systemctl --user status pipewire# Enter the Bluetooth controller
bluetoothctlInside the bluetoothctl prompt, run:
power on
agent on
default-agent
scan on
# Wait until your speaker appears; note its MAC XX:XX:XX:XX:XX:XX
pair XX:XX:XX:XX:XX:XX
trust XX:XX:XX:XX:XX:XX
connect XX:XX:XX:XX:XX:XX
exit
Set as default output and test:
# List PipeWire nodes; note your speaker's Sink ID
wpctl status
# Set default output (replace <sink-id> with the ID you noted)
wpctl set-default <sink-id>
# Test sound output
pw-play /usr/share/sounds/alsa/Front_Center.wav# Plug in the USB mic, then list audio Sources
wpctl status
# Note the Source ID for your USB mic (e.g., 66) and set it as default:
wpctl set-default <source-id>
# Quick record test from the default source
pw-record --rate 44100 --channels 1 test.wav
# Speak ~3 seconds, then Ctrl+C
# Play back through default sink (BT speaker)
pw-play test.wav# Create project
mkdir -p ~/voice-chatbot
cd ~/voice-chatbot
# Python virtual environment
python3 -m venv .venv
source .venv/bin/activate
# Upgrade pip and install packages
pip install --upgrade pip
pip install faster-whisper ollama numpy kokoro
# Optional: thread tuning for Faster-Whisper (shell startup)
echo 'export OMP_NUM_THREADS=4' >> ~/.bashrc
source ~/.bashrc
# Ensure venv is active for the rest of the setup
source .venv/bin/activate# Install Ollama server
curl -fsSL https://ollama.com/install.sh | sh
# Enable and start the Ollama daemon
sudo systemctl enable --now ollama
# Download a small, fast model
ollama pull gemma3:270m
# Verify it responds
ollama run gemma3:270m "Say hello"# Create the chatbot script (paste in the chatbot.py script from the repo)
nano chatbot.pyFind your mic Source ID via wpctl status, then launch:
# Example: MIC_TARGET=66 (replace with your actual Source ID)
MIC_TARGET=66 python3 chatbot.py# Enable SPI in raspi-config:
sudo raspi-config # Interface Options → SPI → Enable
# Add your user to GPIO/SPI groups
sudo usermod -a -G gpio,spi $USER
# Reboot to apply
sudo rebootAfter reboot, log back in and continue.
# On a Raspberry Pi 5, remove legacy RPi.GPIO first:
sudo apt purge -y python3-rpi.gpio
sudo apt autoremove -y
# Install required packages
sudo apt install -y python3-pip unzip wget git python3-lgpioIf you are not on a Pi 5, you can skip the purge step.
cd ~
wget https://files.waveshare.com/upload/8/8d/LCD_Module_RPI_code.zip
unzip -o LCD_Module_RPI_code.zip# Reuse the existing project venv (create only if it doesn't exist)
cd ~/voice-chatbot
[ -d .venv ] || python3 -m venv .venv
source .venv/bin/activate# Install display-related Python packages into the project venv
pip install pillow numpy spidev gpiozero rpi-lgpio lgpio
# (Optional) Use lgpio pin factory for gpiozero
export GPIOZERO_PIN_FACTORY=lgpio# From inside the venv
cd ~/LCD_Module_RPI_code/RaspberryPi/python/example
python3 1inch28_LCD_test.pycd ~/voice-chatbot
# Create the Bob chat script (paste in the bobchat.py script from the repo)
nano bobchat.py
# Save and exit# Link Waveshare Python lib into the project for imports
ln -s "$HOME/LCD_Module_RPI_code/RaspberryPi/python/lib" "$HOME/voice-chatbot/lib"
# Sprites/images for Bob UI (optional but recommended)
git clone https://github.com/OminousIndustries/Bob_Images.git ~/voice-chatbot/sprites
# Run Bob chat (replace with your mic Source ID)
MIC_TARGET=<source-id> python3 bobchat.py# Devices & routing
wpctl status
wpctl set-default <sink-or-source-id>
# Audio tests
pw-play /usr/share/sounds/alsa/Front_Center.wav
pw-record --rate 44100 --channels 1 test.wav
# Bluetooth pairing (inside bluetoothctl)
power on
agent on
default-agent
scan on
pair XX:XX:XX:XX:XX:XX
trust XX:XX:XX:XX:XX:XX
connect XX:XX:XX:XX:XX:XX
exit
# Core services
sudo systemctl enable --now bluetooth
systemctl --user enable --now pipewire wireplumber pipewire-pulse
sudo loginctl enable-linger $USER
# Ollama
sudo systemctl enable --now ollama
ollama pull gemma3:270m
ollama run gemma3:270m "Say hello"- Replace placeholders like
<sink-id>,<source-id>, andXX:XX:XX:XX:XX:XXwith values fromwpctl statusandbluetoothctl. - Re-activate the venv in new shells:
source ~/voice-chatbot/.venv/bin/activate. - On Pi 5, prefer
lgpio/rpi-lgpio; legacyRPi.GPIOis removed above to avoid conflicts.