An Interactive E-ink Picture Frame (Wall-E)

This project brings a "brain" to an interactive picture frame (nicknamed Wall-E), powered by the Google Gemini API. It captures voice commands, processes them using Gemini's multimodal capabilities, and displays generated art on a Pimoroni Inky Impression e-ink display.

Full build video on YouTube

---

✨ Features

• Voice-to-Art: Record audio prompts that Gemini transcribes and uses to generate unique images.
• Multimodal AI: Powered by gemini-3.1-flash-image-preview for advanced multimodal generation.
• E-ink Integration: Automatically resizes and optimizes images for the Pimoroni Inky Impression 7.3" Spectra.
• Audio Feedback: Features a custom soundboard for system status (activation, processing, success, and errors).
• Event-Driven: Triggered by physical hardware buttons.
• Production-Ready: Includes a systemd service for autonomous operation on startup.

---

🔌 Hardware

• Controller: Raspberry Pi Zero 2W
• Display: Pimoroni Inky Impression 7.3" (600x448) Spectra e-ink
• Audio Input: USB Microphone
• Audio Output: MAX98357 I2S Class-D Mono Amp
• Enclosure: 3D-Printed 7.3" Frame (Remixed by TrisM)

Amp Wiring (Raspberry Pi)

Amp Pin	Raspberry Pi Pin
Vin	5V Power
GND	Ground
DIN	GPIO 21
BCLK	GPIO 18
LRCLK	GPIO 19

---

🛠️ Setup (Mac Iteration)

1. Dependencies: Managed via uv. Run uv sync to install everything.
2. API Key:
   • Copy .env.example to .env.
   • Add your GEMINI_API_KEY from Google AI Studio.
3. Run:

   uv run labs/brain_mac.py

---

🍓 Raspberry Pi Deployment

1. Prerequisites

Ensure you have uv installed on your Raspberry Pi:

curl -LsSf https://astral.sh/uv/install.sh | sh
source $HOME/.local/bin/env

2. Transfer the Code

Clone this repository to your Pi:

git clone https://github.com/lahirumaramba/wall-e.git
cd wall-e

3. Verify Hardware

Check that your USB Microphone and Inky Impression display are detected:

arecord -l  # Verify your USB Microphone is listed
# Inky Impression: Detected automatically via the Python script

4. Setup Environment

Install core and Pi-specific dependencies:

cp .env.example .env
nano .env # Configure your API key and hardware settings
uv sync --extra pi

5. Run the Brain

uv run brain.py

Press Button A on the Inky Impression to trigger a recording.

---

🎨 Inky Impression Support

• Button Mapping: Uses GPIO 5 (Button A) for activation.
• Auto-Detection: Uses inky.auto() for seamless setup.
• Saturation: Fine-tune via INKY_SATURATION in .env.

---

🚀 Auto-Start (Systemd)

To run the project automatically on boot:

1. Edit the service file: Open wall_e.service and replace <YOUR_USERNAME> and <YOUR_PROJECT_PATH> with your actual details.
2. Install the service:

   sudo cp wall_e.service /etc/systemd/system/wall-e.service
   sudo systemctl daemon-reload
   sudo systemctl enable wall-e.service
   sudo systemctl start wall-e.service

---

🧪 Experiments (Labs)

The labs/ directory contains secondary, experimental scripts:

• labs/brain_mac.py: A version of the brain optimized for local Mac testing without e-ink hardware.

---

🙏 Credits & Acknowledgments

• Sound Effects: All sound files used in this project are sourced from freesound.org.
• 3D-Printed Enclosure: Inky Impression (2025 Edition) 7.3" Frame Remixed by TrisM.
• Hardware: Built for the Pimoroni Inky Impression series.