TJBot

TJBot is an open-source, DIY maker kit that connects a Raspberry Pi to IBM Watson services for hands-on AI and embodied cognition experiments.

TJBot (named for Thomas J. Watson) is an open-source AI maker kit, developed at IBM Research in 2016 to explore cognitive object design. The platform leverages a Raspberry Pi 3 to integrate physical components (LEDs, servo, microphone) with IBM Watson developer services: Speech-to-Text, Tone Analyzer, and Visual Recognition. Users build the robot's body—often from laser-cut cardboard or 3D-printed files—and program it using 'recipes' in Node.js or Python. This allows the bot to perform specific cognitive tasks, such as waving an arm, analyzing emotional tone in text, or having a conversation, demonstrating the concept of embedding AI into everyday objects.

https://github.com/ibmtjbot/tjbot