gpTea (2024)

intergenerational storytelling through LLM tea making


Course: Designing Intelligence, Spring 2024
Instructor: Marcelo Coelho
Project Duration: 4 Weeks
Project by: Kevin Tang, Kelly Fang









Introduction:


“gpTea” is a tea set that preserves and encourages the sharing of untold stories between separated friends and loved ones through generative AI.  It enhances the ritual of tea drinking by turning each cup into a shared, interactive storytelling experience, encouraging a slower, more reflective form of communication.


















Hardware

gpTea’s hardware consists of three parts, the cup, the tray, and the teapot. The cup features a 1.28 inch round display, ESP32 microcontroller, microphone, and 1 liPo battery. The body of the cup has a magnifying lens built into its geometry to blend the display into the liquid content once the cup is filled. The tray houses a speaker, a stepper motor that manipulates the teapot’s rotation via a timing belt, and a rotary encoder equipped with 2 haptic motors for user feedback and one hall effect sensor for detecting the presence of the cup. All the electronics housed in the tray are connected to an arduino nano microcontroller. 



Interaction Design:
Cup as Universal Touch Point


System Loading animation
AI speaking indicator



Hall Effect detecting when the cup has been lifted.

Displaying blinking red audio recording light






Rotary Dial






Generated Photograph Display



Prototype Iterations












1st Prototype

Functioning cup display

LED tray





2nd Prototype


Full works-like prototype

Motor and Timing Belt Testing


Teapot integration Testing





3rd Prototype

Motor and Stem Integrated with Tray

All Prototypes
Rotary Encoder Testing





4th Prototype


New Teapot with Turned Bearing Pin
Teapot Mounted on Stem

Pour Test




Magnifying Lens






Software Overview


gpTea interfaces with a full-stack application consisting of a React frontend and Express server through both wireless and serial port connection. The application is connected to a MongoDB database to store and retrieve stories. Additionally, the application is configured with API endpoints which the teacup continuously and wirelessly fetches images and effects from.
Diagram of gpTea’s two settings




Story Collecting (State diagram + prompt engineering)



Setting1: State diagram of gpTea’s story collecting setting







Story Telling (State diagram + prompt engineering)




Setting 2: State diagram of gpTea’s story-telling setting

Chat completion prompts used by setting 2

 




Balancing Ritual with Technology