Numtales

🏁 Inspiration

NumTales was created for the MongoDB Challenge as a playful experiment in blending public art datasets with AI-driven storytelling to ignite curiosity in young learners.

✨ What it does

NumTales makes math feel like a storybook adventure for kindergarteners. It wraps number concepts in colorful tales filled with charming characters and vibrant visuals. AI finds patterns in numbers and turns them into personalized stories that make learning feel like play. MongoDB Atlas Search, including vector search, helps surface the perfect mix of content at just the right moment. Google APIs add real-world imagery and sound to keep things lively, while the Metropolitan Museum of Art’s public collection brings art into the mix, helping kids find shapes in sculptures and patterns in paintings. Every math problem becomes a joyful discovery.

🛠️ How it was built

NumTales runs on an Angular front end with a Python Flask back end. It uses MongoDB’s vector search to handle content matching, and Google Cloud Run to deploy and scale the application with ease.

🏆 Accomplishments

Built a working prototype during the hackathon that combines public data, AI storytelling, and advanced search tools.

🧠 Lessons learned

Learned how to use embeddings to make data more meaningful and explored the full potential of MongoDB Atlas Search with vector capabilities. Gained practical experience deploying applications on Google Cloud Run, making it easy to scale and manage backend services.

🔮 What’s next

Plans include improving the quality of AI-generated stories, adding more multimedia, and expanding the dataset library to create even more learning adventures.

🏗️ Built With

• Angular
  
• Python Flask
  
• MongoDB Atlas (including vector search)
  
• Google API integrations
  
• The Metropolitan Museum of Art public dataset
  

Built With

• angular.js
• docker
• mongodb
• python
• typescript