Alpha

Revolutionize learning by integrating AI-generated videos, Desmos graphing, and other visual tools like Mermaid to transform complex concepts into clear, engaging insights that help empower learning.

Comment

Inspiration

We were inspired to create a better alternative to the traditional study mashsup of 10 different websites - a unified sandbox, Alpha, that brings together all the diverse tools STEM majors rely on. Instead of hunting for obscure videos or resources, Alpha directs you in real time to the tool that perfectly matches your needs while crafting personalized, short-form learning videos to simplify and clarify complex concepts. It knows what you want from your chat and helps you get it with simple tools in the shortest time possible.

What It Does

Alpha is an AI-powered assistant that seamlessly integrates dynamic Python video scripting, interactive Desmos graphing, and smart visual tools like Mermaid charts. It understands your queries and intelligently directs you to the right tool for each task, whether you're visualizing math equations, generating explanatory videos, or translating natural language content into engaging, educational formats. There are also additional UI comfort features like notetaking and uploading files so that the AI can help you more specifically and also so that you can remember what you need to for your studying.

How We Built It

The project is built with a clear separation between frontend and backend systems. Our frontend, developed with React and Vite, provides a modern, responsive interface, while Auth0 secures user authentication and Cloudflare tunneling ensures smooth deployment. Videos and audio files are stored in AWS S3 buckets, and our backend integrates with APIs such as GPT, Gemini, Desmos, Mermaid to deliver dynamic content. The codebase is organized into modular components like ChatInterface, MathStudio, and DesmosStudio, each designed to handle specific tasks — by parsing the input language from the user into LaTeX-controlled Desmos expressions and to render markdown and mermaid code to generate an interactive and seamless user experience for the student.

Challenges We Ran Into

During development, we encountered several challenges. Google Cloud credits were unreliable, which affected our access to the Gemini API, and the free tier of Gemini struggled with accuracy. This led us to invest a lot of time and effort in prompt engineering and debugging. Additionally, ensuring that user inputs were correctly transformed—for example, converting complex mathematical expressions into precise LaTeX for Desmos or aligning API requirements. Finally, deployment was also particularly taxing because of how cloudlfare tunneling and the server building was working with Docker and the different Python, React, and Vite integrations.

Accomplishments That We're Proud Of

  • End-to-End Integration: Successfully created and deployed a product that integrates multiple tools seamlessly.
  • Auth0 Integration: Implemented secure user authentication.
  • Dynamic Graphing: Able to convert user input into actual mathematical expressions and Desmos graphs.
  • Video & Audio Processing: Leveraged Python scripting and AWS S3 for on-demand video and MP3 content generation. The S3 buckets are also assigned as per the Auth0 logins to ensure your generated videos don't go anywhere!
  • Visual Tools: Integrated Mermaid to provide users with interactive diagram and chart creations.
  • Notes and PDF Features: Upload PDFs to personalize Alpha responses and make them specific to your task. Add your own notes to the chat to ensure you have what you need for your learning!
  • Math Studio: Work with Desmos alongside your personal chat in Alpha.

What We Learned

We deepened our understanding of backend-frontend integration, deployment with cloudflare, and debugging across multiple API integrations. We learned the importance of constant testing, debugging issues and of handling dependencies with LaTeX processing and visual data rendering. The time constraints also pushed us to constantly work on training the AI model with different prompting files and constant training.

What's Next for Alpha

Moving forward, we're looking to improve Alpha even by integrating more advanced AI models, such as those from Anthropic or Claude, and enhancing our GPT integrations for improved accuracy and performance. Our goal is to continue refining the user experience with more personalized and precise educational content, pushing the boundaries of interactive learning in a continuously evolving tech landscape.

Share this project:

Updates