Inspiration
Education technology is becoming more digital, but not always more engaging or accessible for kids. Most tools still rely on screens, text, and passive learning, yet children remember experiences far more than worksheets.
We wanted to explore: what if learning wasn’t just something kids looked at, but something they actually experienced?
That question led to StoryRover, a system where children learn through movement, storytelling, and decision-making. Instead of reading about financial concepts, kids make choices that drive a story, and those choices are reflected through physical robot movement and AI-guided narration. The AI acts as a guide, turning lessons into interactive adventures rather than static information.
What it does
StoryRover creates an AI-powered interactive learning loop combining a tablet dashboard and a physical robot.
A learning session looks like this:
1. AI generates a scenario
“You earned 10 dollars helping a neighbor. What should you do?”
2. The child makes a choice on the tablet
- Spend
- Save
- Invest
3. The robot moves physically
It drives to a colored floor zone representing the selected choice.
4. The robot speaks the outcome
An expressive AI voice narrates what happens next in the story.
5. The child sees the result
- Balance updates
- Story consequences unfold visually
6. The loop continues
Each scenario is different, creating an ongoing interactive story.
How we built it
Frontend: React + TypeScript + Tailwind
Backend: NestJS services for game logic, AI, hardware
AI: OpenAI GPT for scenario & consequence generation
Hardware: Raspberry Pi, Arduino motors, ElevenLabs voice
Challenges we ran into
Connecting the Raspberry Pi reliably
- Maintaining stable communication between the laptop, Raspberry Pi, and Arduino in real time was difficult. Network drops or timing delays could interrupt the robot’s movement or voice output, so we had to carefully manage WiFi setup and command handling.
Keeping AI responses age-appropriate and clear
- Because the AI generates stories dynamically, we had to design prompts and structured output rules to ensure explanations stayed simple, educational, and suitable for kids aged 7–12.
Accomplishments that we're proud of
- Fully working phygital accessibility learning system
- Real-time AI storytelling
- Seamless AI + hardware integration
What we learned
- Integrating AI generation with real-time physical hardware requires careful latency handling and fallback logic.
- Voice AI feels more “alive” when responses are short, expressive, and designed for children, not just raw text output.
- Connecting a tablet dashboard, backend server, and Raspberry Pi robot over local WiFi taught us a lot about device communication and reliability.
- Physical interactions introduce real-world constraints (battery, movement accuracy, delays) that software-only projects don’t face.
What's next for StoryRover: Money Quest
Expanding Learning Content
- Add more financial scenarios covering budgeting, delayed gratification, needs vs wants, and risk.
- Introduce difficulty levels by age group to grow with the child.
- Support multiple languages to make the platform more inclusive.
Smarter AI Personalization
- Adapt scenarios based on a child’s past choices and learning patterns.
- Provide progress tracking and learning insights for parents and teachers.
- Adjust story complexity and vocabulary to match reading level.
Hardware Improvements
- Improve robot movement precision and speed.
- Add lights, sounds, and emotional expressions to make the robot more engaging.
- Make the system easier to set up in classrooms.
Classroom & Home Use
- Develop a teacher dashboard for group learning activities.
- Create lesson plans aligned with school financial literacy standards.
- Design a portable kit for libraries, schools, and community centers.
Long-Term Vision
StoryRover becomes a platform where AI-powered storytelling and physical interaction help children learn complex life skills, not just financial literacy, but decision-making, responsibility, and critical thinking.
Built With
- arduino
- elevenlabs
- nestjs
- openai
- python
- raspberrypi
- react
- tailwindcss
- typescript


Log in or sign up for Devpost to join the conversation.