SparkyAI

Inspiration

As a parent, I was amazed by the power of AI like Gemini but concerned about its safety for children. This sparked my idea for SparkyAI: a "walled garden" for AI learning where kids could explore educational topics with a safe, encouraging, and engaging AI companion they could trust.

What it does

SparkyAI is a safe AI tutor for kids that answers educational questions using a curated knowledge base stored on MongoDB, featuring multiple friendly character personas to make learning fun and interactive.

How I built it

With no coding experience, I used Firebase Studio's AI to generate a Next.js app from my own HTML designs, connected it to a MongoDB database I curated, and implemented a safe RAG chat system using the Gemini API.

Challenges I ran into

My biggest challenge was making the AI use my curated MongoDB content instead of its general knowledge, which required debugging the backend logic and strengthening the prompts sent to the Gemini API.

Accomplishments that I am proud of

I'm proud of building a fully functional and deployed AI application from scratch with no prior coding experience, successfully implementing a safe, content-controlled chat system with unique AI personas.

What I learned

I learned that building great AI applications is less about coding and more about meticulous prompting, systematic debugging, and providing clear instructions for the AI to follow to achieve a specific, safe outcome.

What's next for SparkyAI

The next steps are to build out the parental dashboard for monitoring progress, expand the MongoDB knowledge base with more subjects, and introduce new interactive content types like quizzes and games.

Built With

• aistudio
• firebasestudio
• geminiapi
• googlecloudai
• mongodb
• next.js
• tailwindcss