AstroNotes

AstroNotes is a generative note-taking platform that listens to lectures and conversations in real time. It transforms audio into structured, interactive study tools like dynamic mind maps, flashcards, quizzes, summaries, slideshows, and even a chatbot you can ask questions to.

Inspiration

We were inspired after noticing students recording lectures, and then utilizing AI study tool in separate platforms. We wanted to build something that not only captures information, but actively helps process and understand it.

What it does

AstroNotes records audio and transcribes it in real time. From there, it uses AI to:

Build a dynamic, evolving mind map of key topics
Generate flashcards for studying
Create quizzes to test understanding
Produce concise summaries
Turn content into slideshow-style presentations
Provide a chatbot interface to ask questions about the material

All of this happens automatically, turning passive listening into active learning.

How we built it

We built AstroNotes using a full-stack architecture:

Frontend: Records audio and provides an interactive UI for mind maps and study tools
Backend: Handles transcription, processing, and AI generation
AI Pipeline: Converts raw transcripts into structured knowledge using embeddings and language models

We focused heavily on real-time processing and keeping everything responsive as new information comes in.

Challenges we ran into

Handling real-time audio streaming and ensuring reliable transcription
Converting unstructured lecture data into meaningful, organized outputs
Managing token usage and performance when processing large transcripts
Designing a system that updates dynamically without overwhelming the user

Accomplishments that we're proud of

Building a fully automated pipeline from audio to structured study tools
Creating a dynamic mind map that evolves in real time
Integrating multiple AI-powered features into a single cohesive platform
Making the experience intuitive and useful for actual students

What we learned

Real-time cloud systems introduce a lot of complexity compared to local processing
Structuring information is just as important as generating it
UI/UX plays a huge role in how useful AI tools actually feel
Performance and cost optimization are critical when working with large models