DIY Meckie — Build Your Own AI Companion

🛒 What you need

Everything is off-the-shelf. No custom PCBs, no soldering required.

🖥️ Raspberry Pi 3B+ or newer

~$35-55

The brain. Pi 3B+ works great. Pi 5 is faster but not required.

📷 USB Webcam

~$15-30

Any USB webcam with a mic. Logitech C260/C270 are solid budget picks.

🔊 USB Speaker or 3.5mm Speaker

~$10-25

USB speaker for Pi 3B/3B+. USB, 3.5mm aux, or Bluetooth on Pi 4/5.

💾 MicroSD Card (16GB+)

~$8

For the OS. 32GB recommended.

🔌 Power Supply

~$10

Micro USB 5V 2.5A for Pi 3B/3B+. USB-C 5V 5A for Pi 4/5.

🎯 Pan/Tilt Servo Kit (optional)

~$15-20

2x SG90 servos + pan/tilt bracket. Lets your Meckie track faces and look around.

🎛️ PCA9685 Servo Board (optional)

~$5

16-channel PWM driver. Controls the servos from the Pi's I2C bus.

🤖 3D Printed Body (optional)

Free STLs

Print your own enclosure. STL files coming soon — or design your own!

Total cost (without Pi)

~$45-75

Most people already have a webcam and speaker lying around

⚡ Setup in 5 minutes

1

Flash Raspberry Pi OS

Download Raspberry Pi Imager and flash Raspberry Pi OS Lite (64-bit) to your SD card. Enable SSH and set your WiFi in the imager settings.

2

Install the Meckie client

SSH into your Pi and run one command:

cd ~ && git clone https://github.com/rangerchaz/droid-pi.git droid && cd droid && bash install.sh

3

Sign up & get your token

Create a free account at meckie.ai. Go to your dashboard → Hardware → generate a device token.

4

Paste your token

nano ~/droid/config.json
# Paste your device token, save, and exit

5

Start your Meckie

sudo systemctl start droid

That's it. Your Meckie will say hello, recognize your face, learn your voice, and start remembering.

🧩 What it can do

Face recognition — knows who's in the room
Voice identification — recognizes you by your voice, even in a crowd
Persistent memory — remembers your conversations, preferences, and the people in your life
Skills — weather, music, reminders, shopping lists, web search, and more
Meeting mode — transcribes meetings with speaker attribution and action items
Song ID — "what song is this?" and it tells you
Multi-language translation — real-time translator mode
Head tracking — add servos and it follows you around the room (optional)
Fully hackable — it's Python on a Pi. Go wild.

🖨️ 3D Printable Body

PIP-inspired droid body with snap-fit assembly. Camera in the head, pan/tilt servo mount, docking station.

STL files coming soon. Join the Discord to get notified.

Ready to build?

Start with a free trial. No credit card required.

Get Started Free

View source on GitHub →

Build your own Meckie