rabelus-lab-agi

• 

  DIDATA - Pedagogical AI Architect - by Rabelus Lab

Inspiration

My inspiration came from the frustration with passive learning. I wanted to transform the reading of dense texts and complex books — often a solitary journey — into a dynamic and interactive conversation. The idea was simple: what if we could "talk to knowledge"?

What I built

I built Didata, an AI Pedagogical Architect. It is not just another chatbot. Didata consumes raw, static content (such as a PDF or a long text) and reconstructs it as a structured course, with logical modules and sequential lessons. The key differentiator is its "Socratic Tutor": an AI agent you can talk to by voice in real time, powered by the Gemini Live API. You can interrupt it, ask questions, and be guided through the content — turning studying into an active dialogue.

How I built it

The heart of the project is the Gemini Live API, which enables fluid, real-time voice interaction. For content structuring, I used Gemini Pro to analyze raw material and organize it into a pedagogical framework. The interface was built with React, TypeScript, and Vite, focusing on a clean and responsive user experience. To ensure privacy and speed, all user data is stored locally in the browser via IndexedDB.

Challenges I faced

The main challenge was twofold: technical and conceptual. On the technical side, optimizing the architecture to minimize voice conversation latency was critical to making the interaction feel natural. On the conceptual side, the challenge was designing the prompt and logic of the "Socratic Tutor" so it would act as a true guide — asking questions, provoking reasoning — rather than simply delivering ready-made answers.## Inspiration

What it does

How we built it

Challenges we ran into

Accomplishments that we're proud of

What we learned

What's next for rabelus-lab-agi

Built With

• gemini-live-api
• gemini-pro
• google-ai-sdk
• google-cloud
• react
• tailwind-css
• typescript
• vite