Lecturix

Transform your lecture notes into dynamic video!

Comment

Inspiration

Education has struggled to keep pace with the rapid advancement of technology. While tech continues to evolve at lightning speed, the education system remains largely unchanged, creating a vast gap and a significant opportunity for transformation. Many professors are unable to record lecture videos, whether due to time constraints or institutional policies.

Our inspiration comes directly from what we’ve seen in our classrooms—professors often express a desire to do more. They want to hold more office hours, offer extra resources, and provide greater support, but time just isn’t on their side. To bridge this gap and empower both educators and students, we created Lecturix—a tool designed to bring meaningful change to the learning experience.

What it does

Lecturix revolutionizes the way professors offer supplemental materials by turning lecture notes into dynamic, personalized videos—without adding to their workload. Professors simply upload their lecture notes to our platform, and Lecturix does the rest. Using Gemini and ChatGPT 4 Turbo, the platform automatically generates a video based on the notes that the professor uploads. It then integrates Cartesia to create realistic voice overs, synchronizing the audio perfectly with the video. Professors can even personalize these videos with their own voice, enabling them to maintain their unique teaching style without spending extra time recording.

Professors can organize their videos into specific folders tied to their courses, ensuring that students have easy access to the relevant content. Each folder is labeled with a class name and code to maintain organization.

For security and ease of use, both professors and students create accounts to access features—professors for uploading notes and students for watching videos. All data, including videos and user information, is securely stored using SingleStore, ensuring a seamless and reliable experience for everyone involved.

Lecturix empowers professors to provide personalized, on-demand learning tools without added effort, enhancing the educational experience for students.

How we built it

Initial Planning: We started by using Figma to visualize and design our ideas, mapping out the user interface and overall user experience. We also created a pipeline for the lecture notes to video animation aspect.

Frontend: The frontend of the application was built using React, allowing us to create an interactive and responsive interface for both professors and students.

Backend: We integrated Gemini and ChatGPT 4 Turbo to automatically generate video animations based on the lecture notes uploaded by professors. For voiceover, we used Cartesia to create realistic, synchronized audio for the videos. We utilized SingleStore to securely store user information and easily retrieve the videos for on-demand access. Flask was used to connect the frontend and backend, ensuring smooth communication between the two, and we hosted the website on Render.

Challenges we ran into

During the development process, we faced challenges with text to audio and video. We didn't exactly know how we could achieve this so we thoroughly researched different voice APIs and Cartesia seemed like the optimal choice.

Another obstacle we encountered involved GPT-4 Turbo, which didn't always generate relevant animations to accompany the voiceovers. To prevent displaying faulty or irrelevant animations, we opted to provide the script alongside the AI narration in our lectures, allowing students to follow along by reading. Additionally, GPT-4 Turbo sometimes generated scripts with complex math syntax and symbols, making it difficult to understand through the AI voiceover. To address this, we used prompt engineering techniques to ensure the lecture content was presented in a more accessible and comprehensible way for the students.

Accomplishments that we're proud of

Creating a User-Friendly Interface: Designing a visually appealing and functional user interface required multiple iterations, experimenting with different color schemes and component layouts to achieve the best result.

Adapting to Unfamiliar Technologies: Learning to work with new technologies, such as Flask, was challenging. Additionally, merging multiple individual's code into one cohesive project required careful planning and troubleshooting to ensure everything integrated smoothly.

What we learned

Collaboration and Problem-Solving: The challenges we encountered strengthened our teamwork and problem-solving skills. We learned how to efficiently debug issues and communicate solutions, which was essential for keeping the project on track.

What's next for Lecturix

We’re exploring the option of allowing students to upload their own notes, enabling them to create personalized learning materials and access additional resources for deeper, tailored learning. We will also implement fine-tuned versions of Gemini and GPT-4 Turbo in our pipeline for more accurate, bug-free, and nuanced diagram generation.

Built With

Share this project:

Updates