From an exciting array of low-cost hardware to more plug-and-play training modules, it’s easier than ever for makers to incorporate machine learning into personal projects. In the new issue of Make: (Vol. 77), we take a look at new tools you can use to train AI modules right now like TensorFlow, Edge Impulse, and Lobe, and show how to use frameworks like TensorFlow Lite to squeeze trained AI onto a Raspberry Pi or Arduino.

Our cover story details how to make an AI-empowered artificial nose that can sort coffee from tea, or identify whatever else you train it to smell. Next in this section we show: how to build a teddy bear that listens and Tweet-shames you when it hears a swear word, how to make a trash-sorting camera using Lobe and Raspberry Pi, and how to monitor your 3D printer for failed prints with an AI watchdog. 

Plus, get the inside scoop on Tesla coil-powered band ArcAttack’s new mini musical coil kit.

Also find 23 projects inside, including:

• Track meteors entering the atmosphere with a Raspberry Pi
• Modify a vintage iPod to stream music from Spotify
• Make an animated light-up skirt with 120 hidden full-color LEDs that respond to your movements
• Add some arctic ambience to your home with a quick Arduino-powered Aurora Borealis light setup
• And much more!