The Pomodoro Bot V2 is a significant upgrade from its predecessor, featuring smart voice interaction, enhanced design, and streamlined hardware. This version goes beyond just tracking focus sessions—it can now respond to voice commands and provide real-time feedback. With LLM integration, the bot operates completely hands-free, eliminating the need for physical buttons.

To complement these advancements, we have redesigned the bot with a more compact form factor, custom PCB, and an adorable new look with ears.

The first iteration of the Pomodoro Bot featured a simple display for tracking focus sessions and physical buttons for manual control. While it served its purpose well, we aimed to take productivity to the next level with a more intelligent and interactive experience.

If you haven't seen the original version in action, you can check out the V1 documentation here.

Let's dive in and see what's new in Pomodoro Bot V2!

Pomodoro Bot v2

What's New in v2?

Watch The Video

Before getting into the build, let's start by seeing our Pomodoro Bot V2 in action!

Prerequisites

One of the most significant upgrades in Pomodoro Bot v2 is the audio system, enabling natural voice interaction. We focused on compact and efficient components for both audio input and output, ensuring seamless integration into the design with minimal modifications.

Audio Input: USB Microphone

For voice commands and interaction, we selected a small USB microphone that fits perfectly with just a minor tweak to the 3D model. This allows for clear voice detection without requiring additional circuitry.

USB Microphone

Audio Output: Speaker & Amplifier

For audio feedback, we chose a compact 4Ω, 2.5W speaker, powered by a PAM8403 amplifier module, ensuring clear and crisp sound.

4Ω, 2.5W Speaker

PAM8403 Amplifier

Audio Routing: A Creative Solution

Since the Raspberry Pi 5 lacks a built-in audio output, we needed an alternative approach:

  1. Instead of using an external sound card, we routed the audio signal through the display's 3.5mm jack.
  2. This signal was fed into the PAM8403 amplifier, which then powered the speaker.

Power & Noise Reduction

Testing the Audio System

To verify functionality, we ran basic recording and playback tests using the following commands:

# Record audio
arecord -D plughw:1,0 -f cd -d 5 test.wav  

# Play recorded audio
aplay test.wav

With these adjustments, Pomodoro Bot v2 now has a voice, making interactions more natural and engaging!

With the audio setup complete, the next step is integrating VIAM and Local LLM to enable real-time voice interactions. This allows the Pomodoro Bot to process speech, understand queries, and respond intelligently.

Adding Local LLM to the Bot

VIAM's Local-LLM module enables running offline AI models without relying on the cloud. It supports TinyLlama, Deepseek R1, and other models using llama.cpp.

Steps to Integrate Local LLM

  1. Navigate to the CONFIGURE tab in VIAM.
  2. Search for LOCAL-LLM under services and add it to the bot.
  3. To use a custom model, you'll need:
    • LLM-File: The actual model file in GGUF format.
    • LLM-Repo: The repository ID containing the model.

Using Deepseek R1 Model

  1. Search for Deepseek R1 GGUF on Hugging Face.
    HF Search
  2. Download the GGUF file to the Raspberry Pi (this is the LLM-File).
    GGUF Download
  3. Note down the repository ID (this is the LLM-Repo). GGUF Repository ID
  4. Back in VIAM's CONFIGURE tab, set:
    • LLM-File: Path to the downloaded GGUF file.
    • LLM-Repo: Repository ID of the model.
      LLM Config
  5. Save the configuration, and the bot is now powered by an offline LLM!

With LLM integration, the Pomodoro Bot can now listen, understand, and respond in real time, making it a truly hands-free experience.

To enable hands-free interaction, we're integrating a wake word detector into the Pomodoro bot using PicoVoice Porcupine Wake Word Engine. Porcupine is:

Setting Up the Wake Word

Access PicoVoice Console

Choosing or Creating a Wake Word

Downloading the Wake Word Model

Getting the PicoVoice API Key

Next, we'll configure the wake word engine in our code. Here's a sample code to test whether your wake word detection is working fine or not.

#wake_word_test.py

import pvporcupine
import pyaudio
import struct

# Replace with your actual access key and wake word model file path
ACCESS_KEY = "YOUR_ACCESS_KEY"
WAKE_WORD_MODEL = "path/to/your_wakeword.ppn"

# Initialize Porcupine wake word engine
porcupine = pvporcupine.create(access_key=ACCESS_KEY, keyword_paths=[WAKE_WORD_MODEL])

# Configure audio stream
pa = pyaudio.PyAudio()
stream = pa.open(
   rate=porcupine.sample_rate,
   channels=1,
   format=pyaudio.paInt16,
   input=True,
   frames_per_buffer=porcupine.frame_length
)

print("Listening for wake word...")

try:
   while True:
      pcm = stream.read(porcupine.frame_length, exception_on_overflow=False)
      pcm = struct.unpack_from("h" * porcupine.frame_length, pcm)
      result = porcupine.process(pcm)
      
      if result >= 0:
            print("Wake word detected! Activating voice interaction...")
except KeyboardInterrupt:
   print("Stopping wake word detection...")
finally:
   stream.stop_stream()
   stream.close()
   pa.terminate()
   porcupine.delete()

To run the above code install the required libraries using following commands

sudo apt-get install portaudio19-dev
pip install pvporcupine pyaudio

Now run the code as

python3 wake_word_test.py

And just like that, our Pomodoro bot is now intelligent, interactive, and fully voice-enabled!

To maintain a neat and clutter-free assembly, we designed a custom two-layer PCB using EasyEDA. The PCB layout was optimized to facilitate seamless connections between components, ensuring a compact and efficient design.

Features

Design Process

  1. Designed the PCB using EasyEDA. PCB Design
  2. Exported Gerber files for manufacturing. Export As Gerber

After the fabrication process was completed, we received our custom PCBs, precisely manufactured according to our design specifications.

Fabricated PCBs

This custom PCB simplifies the wiring process, enhances durability, and improves the overall aesthetics of the Pomodoro bot. You can get your own PCBs using this Gerber File.

In addition to designing the custom PCB, we also gave the Pomodoro bot's case a complete makeover. The updated design enhances both aesthetics and functionality, creating a more charming and efficient look.

3D Design

Below are the key features of the redesigned case:

3D Printing the New Design

Once the redesign was finalized, we moved on to preparing the 3D printing files and began printing the individual parts. The choice of materials and colors was carefully considered to complement the bot's appearance:

This redesign elevates the bot's functionality and personality, creating a more cohesive and delightful user experience.

3D Printed Parts

Find the 3D Files here.

Front Portion Assembly

  1. Speaker Installation
    • Place the speakers onto the speaker grill and ensure proper alignment. Speakers

      Speakers Assembled
  2. PCB Assembly
    • Solder all necessary components onto the PCB. PCB Soldering
    • Use female headers for the ADS1115 module to allow easy replacement. ADS1115
    • Directly solder the PAM8403 module for a secure connection. Assembled PCB
  3. Mounting the PCB on the Raspberry Pi
    • Attach the 3.5mm audio jack to the PCB. 3.5 MM Jack To RPI
    • Stack the assembled PCB onto the Raspberry Pi. PCB On RPI
    • Insert the entire unit into the front portion of the case. PCB In Case
    • Connect the speakers to the PCB. Speakers To PCB

Back Portion Assembly

  1. Sensor and USB Module Installation
    • Solder JST connectors to both sensors. JST To Sensors
    • Attach the sensors and USB module to the back side of the case. Back Portion
  2. Final Wiring
    • Connect the wiring between the back portion and the front portion, completing the internal assembly. Back To Body

Final Assembly

  1. Securing the Outer Case
    • Glue the front ring and ear portions to the Pomodoro bot's body. Ears

      Front Ring

      Assembled Pomo

Assembly process fully completed! The Pomodoro Bot is now ready for use.

1. Power Up the Raspberry Pi

Once the assembly process is complete, power up your Raspberry Pi. You can do this by connecting the power supply to the Raspberry Pi's power jack.

2. Verify Functionality

After powering up, take a moment to verify that everything is functioning as expected. This may involve checking the Raspberry Pi's LED indicators and ensuring the Waveshare display turns on.

3. Download Source Code

The next step is to download the source code for the Pomodoro Bot application on to your Raspberry Pi. You can download the code from the following GitHub repository:

4. Edit the Source Code (main.py)

5. Run the Pomodoro Bot Code

6. See Your Pomodoro Bot Alive!

If everything is configured correctly, the Pomodoro Bot application should launch and your Raspberry Pi will transform into a functional Pomodoro Bot, ready to help you manage your work sessions and boost your productivity.

Assembled Product