Chikki - Virtual Pet 🐾🤖

Inspiration

Learning and engagement improve when interactions feel natural and personalized. Inspired by virtual companions in gaming and AI-driven personalized learning assistants, we created Chikki—an AI-powered desktop pet that listens, responds, and adapts to users in real time.

What it does

Chikki is an interactive AI-powered virtual pet that:

  • 🗣 Listens to user conversations using speech recognition.
  • 💬 Engages in real-time chat with an AI-driven personality.
  • 🔊 Speaks responses aloud using advanced text-to-speech (TTS).
  • 🎭 Expresses emotions & animations based on interactions.
  • 📚 Personalizes learning by adjusting to user interests and pace.
  • 📅 (Future) Integrates with productivity apps to provide task reminders and structured goals.

How we built it

  • Godot 4 & Pygame for animations, UI, and desktop interaction.
  • Python (Flask & Speech Recognition) for handling voice input.
  • Edge TTS for lifelike AI voice responses.
  • Ollama LLM for generating intelligent, adaptive conversations.
  • Retrieval-Augmented Generation (RAG) (Future) for context-aware reminders and goal tracking.
  • HTTP-based API for smooth communication between Godot and the AI backend.

Challenges we ran into

  • 🔧 Managing real-time speech recognition with accurate AI responses.
  • 🎙 Ensuring seamless audio playback without caching issues in Godot.
  • 📡 Optimizing HTTP-based interactions to minimize latency.
  • 🎨 Balancing animations & interactivity to keep Chikki engaging.

Accomplishments that we're proud of

  • ✅ Built a fully interactive AI-powered pet that listens and talks.
  • ✅ Created a dynamic learning companion that adjusts to user input.
  • ✅ Overcame file caching issues for real-time voice responses.
  • ✅ Developed adaptive conversations using an LLM-powered chatbot.

What we learned

  • 🧠 The importance of real-time AI interactions for engagement.
  • 🎭 How animations and personality improve user retention.
  • 🎙 Improved understanding of speech recognition & TTS pipelines.
  • 📡 Optimizing Godot-Python communication via HTTP requests.

What's next for Chikki - Virtual Pet

🌟 Expanding Use Cases

Chikki has applications across various fields:

  • 🧠 Personalized Learning – Adaptive AI tutor that adjusts to user pace.
  • 🎮 Gamification – Interactive challenges & rewards for engaging learning.
  • 🧘 Mood & Stress Management – AI-powered mood tracking & stress reduction.
  • 👥 Productivity & Task Management – Smart AI assistant that organizes schedules and tasks.

🚀 Future Features & Enhancements

  • 📅 Integration with Productivity Apps

    • Connects with Google Calendar, Notion, and task managers to provide smart reminders.
    • Uses Retrieval-Augmented Generation (RAG) to analyze user tasks and create goal-based milestones.
    • Rewards users upon completion of milestones to boost motivation.
  • 🎭 Emotion & Mood Analysis

    • Detects user expressions & speech tone to suggest mood-improving activities.
    • Provides stress-reduction techniques based on detected mood shifts.
  • 📊 Productivity Analytics & Well-being

    • Takes periodic screenshots and analyzes user activity to recommend breaks and focus sessions.
    • Tracks screen time, distractions, and efficiency trends to improve productivity.
    • Suggests mindfulness exercises when excessive stress is detected.
  • 🕹 AI-driven Interactive Challenges

    • Encourages learning and productivity through gamified tasks and rewards.
    • Engages users with adaptive storytelling and task-based missions.

Chikki is more than a virtual pet—it’s a smart, interactive AI assistant that enhances learning, productivity, and well-being! 🐾🚀

Built With

Share this project:

Updates