StimStudy

Generate Study Plan
Review & Refine
Choose Character & Background
Learn & Enjoy!

Inspiration

We asked ourselves, “What if studying could be as stimulating as scrolling through social media?” That question inspired StimStudy — an educational platform that blends AI-generated study plans with character narration and the engagement of entertainment apps to make learning both focused and fun.

We wanted to build a tool that doesn't fight for attention — it works with how students already consume content, turning distractions into learning opportunities.

What it does

StimStudy generates a personalized feed of short form educational content. This is perfect for the student who wishes they could watch a personalized tik tok video explaining the concept they just keep getting stuck on!

For example: “teach me about the Pythagorean Theorem, I can solve for c but not a or b”

Our platform will generate a few short-form pieces of content that describe the concept. Each video is complete with animated slides, and someone such as Peter Griffin, SpongeBob, or Joe Biden explaining the Pythagorean Theorem, with slang familiar and beloved by genZ!

This all happens in an intuitive, easy to understand interface in which the user can refine a study plan, choose the character, and choose the background video. Once they click ‘Generate’ our AI and Video rendering system will get to work concurrently generating these videos.

Once the videos are generated the user is brought to a familiar intuitive scrollable content viewer, and are able to ask additional clarifying questions about the concept to Gemini in the ‘Comments’ section!

How we built it

We used React.js to build a smooth, responsive frontend interface that allows users to input prompts, choose characters, and watch their personalized study videos. For the backend, we built a Flask (Python) server that handles AI generation, video processing, and audio syncing. To power the generative AI experience, we integrated Google's Gemini API, which we used to create personalized study plans, refine user input, and generate educational scripts that are later turned into videos. This combination of technologies helped us deliver a fast, engaging, and intelligent user experience from end to end.

Challenges we ran into

One challenge we faced was combining all the parts — AI content, voice narration, animations, and video — into one smooth experience. We had to make sure the AI-generated content felt fun and engaging, not boring like traditional study tools. It was also tricky to make the whole process run quickly, since generating everything one by one took too long. To fix that, we worked on running tasks in parallel to speed things up and make it more efficient.

Accomplishments that we're proud of

Throughout this project, our team overcame multiple challenges to deliver a truly technically impressive AI video generation pipeline. Here are some key moments that stand out to us: Concurrent Video Compositing Pipeline This video pipeline takes as an input HTML files, audio files, video files, and character sprites and combines them in the following way: Render HTML slides into videos Animate character sprite changes that play dynamically with the slides Combine all slides, character animations, and background videos into a final polished product Do this all concurrently so multiple videos can be created simultaneously!

Smooth front end experience Our front end is a powerful abstraction on top of our video creation engine, allowing users to easily generate and refine study plans, choose the character and background video, and generate the final library of videos in just a few clicks.

Gemini-Powered HTML slide animations We feel we took full advantage of the advanced capabilities of Gemini with our animated slide generation. Gemini 2.5 Flash can generate extremely detailed slides with surprisingly intuitive animations of various concepts.

What we learned

We learned how to use generative AI to create content that’s not just accurate, but also fun and engaging for modern learners. We figured out how to build a fully integrated video pipeline that combines HTML-rendered slides, voiceovers, sprite animations, and background videos into one cohesive and visually appealing product. We gained hands-on experience with parallel processing to speed up video generation and learned how to manage concurrency in a way that keeps the system running smoothly. On the frontend, we focused on building an intuitive user experience that walks users through the process of creating content, from selecting a narrator and background to refining study plans and viewing the final videos. We also learned how to simplify complex technology by hiding all the heavy lifting behind a clean and interactive interface. This project challenged us to think both technically and creatively, helping us build something that is powerful behind the scenes while being easy and enjoyable to use.

What's next for StimStudy

We want to deploy the app to mobile app stores so more students can easily access it.

We plan to let users view and explore content created by others, creating a community of shared learning.

We’ll add a history/log feature so users can revisit their past study plans and video creations.

We want to add more voice characters to give users more fun and personalized narration options.

We also plan to add more background environments and character animations to make the videos even more engaging.

Built With

Submitted to

CruzHacks 2025
- Winner Education Hacks

Created by

I worked on the frontend and also was connecting all of everyone's work with our backend service to make sure it all ran seamlessly when called through HTTP.

Yoshinobu Sono
I worked mostly on the backend where I handled the AI voice generation through FishAudio's service and API. I also worked on using MoviePy to combine all of our smaller parts into a cohesive video which includes the audio, multiple videos, and animations.

Aaditya Jadhav
I contributed to some smaller features on the frontend, but my main focus was on the backend, where I worked on the video generation pipeline. I was responsible for integrating all the individual components our functions produced—like audio, character animation, and background elements—into a cohesive final video output

Harshil Gupta
I worked on creating the gemini-powered animated slides, the HTML rendering engine, and automating the video asset creation process. Also a bit on the front end for video playback

Colton Kirsten