An AI companion for personal growth

Personalized gEnerative Tamagotchi

Personalized technology to support self-directed goals & behavior development.

Try the interactive demo →

PET roster with four personalized companion cards, each showing a pet, its goal, level, health, and experience.
Each PET represents a personal goal, with its own appearance and progress indicators.

How It Works

PET turns a goal you describe in your own words into a companion generated for that goal, then keeps you company while you pursue it.

Describe your goal

Tell PET what you want to work on, in plain language, and pick an art style. Any self-directed goal works — there is no fixed list of categories.

Meet your generated pet

Generation agents build a persona, a plan of concrete actions, and a one-of-a-kind animated companion from your goal and style choice.

Care for it daily

The CareTaker agent chats with you, logs what you did, and turns your progress into your pet’s health, experience, and level.

Motivation

PET studies how an AI pet can help people set meaningful goals, stay motivated, and turn intentions into sustainable actions.

Problem 01

Addressing App Abandonment

How can digital systems prevent users from quickly abandoning self-directed goal trackers after initial setup?

Problem 02

Overcoming Single-Domain Constraints

How can generative AI convert open-ended, user-defined goals into personalized virtual pets rather than relying on fixed, single-domain characters?

Problem 03

Companion Attachment and Identification

How does generating a companion directly from a user’s self-described goals shape their emotional attachment, sense of ownership, and companion identity?

PET in conversationA conversation with LEAV connects everyday sustainable choices with companion responses and progress tracking.
PET chat interface showing LEAV, a personalized companion, responding to a user's sustainable-living actions alongside pet health, experience, and activity tracking.

Research Questions

PET investigates how a personalized AI companion affects goal pursuit, daily engagement, and users’ relationships with their pets.

Issue 01

Goal-Pursuit Outcomes

What are the effects of a personalized virtual pet, combined with an AI-supported goal-planning system, on self-directed goal pursuit?

Issue 02

Companion Experience

What are participants’ experiences with the pet as an AI-generated personal companion during goal support?

Issue 03

Sustained Daily Engagement

How does PET’s care-loop mechanic sustain daily engagement beyond initial conversational novelty?

Issue 04

Identification, Ownership & Attachment

How does generative personalization shape users’ identification with, ownership of, and attachment to their companion?

Publications

Research papers, posters, demonstrations, and project reports related to PET.

Sun, Q. & Hsiao, I. H. (2026, Sept). Hyper-Personalized Virtual Companion: Generative Tamagotchi to Cultivate Personal Goals. In Proceedings of the 37th ACM Conference on Hypertext and Social Media (pp. 46–50)

Sun, Q., Chien, S-Y. & Hsiao, I. H. (2026, Nov). PET: A Personalized gEnerative Tamagotchi-Style Companion for Self-Directed Goal Support. In Proceedings of the 14th International Conference on Human Agent Interaction

Sun, Q., Chien, S-Y., Chung, CF. & Hsiao, I. H. (2027, Jan). A Pet for Every Goal: Personalized gEnerative Tamagotchi-Style Companions for Self-Directed Goal Support. In Proceedings of the 60th Annual Hawaii International Conference on System Sciences, 2027

PET system architectureUser goals and visual preferences guide companion generation; the CareTaker agent supports interaction, tracking, and conversation.
PET architecture: user goals and art-style choices feed Profiler, Strategist, and Designer generation agents; the CareTaker agent supports pet interaction, goal tracking, and chat and logging.

What’s Next

Research Directions

  1. Expanded social interaction: Introduce shared spaces, such as a virtual park, where users can “walk” their PETs, discover goals, meet other users, and showcase their personalized companions.
  2. Post-goal companion lifecycle: Create meaningful experiences after a user completes a goal, such as holding a commemorative farewell ceremony or allowing the PET to evolve into a ghost-like form inspired by Obaketchi.
  3. Immersive PET: Imagine your AI companion is not just a companion, but also your travel guide, your workout buddy and so much more…so they go wherever you go in the immersive setting.

About

Personalized gEnerative Tamagotchi (PET) project builds novel multi-agentic AI architecture to create personalized AI pets designed to support users to pursue any of their self-directed goals. PET explores how personalization, companionship, and generative AI can make behavior interventions more engaging and supportive.

This website documents the motivation, research questions, development process, and outcomes of the project.

Team

Qiming Sun

Lead Developer & Researcher
PhD Candidate
Santa Clara University
Currently on the job market

Toby Chen

Undergraduate Research Assistant
Santa Clara University

Dr. Sharon Hsiao

Principal Investigator
Associate Professor
Santa Clara University

ihsiao (at) scu (dot) edu

Collaborators

Dr. Shih-Yi Chien

Associate Professor
National Chengchi University (NCCU)

Dr. Christina Chung

Associate Professor
University of California, Santa Cruz (UCSC)