Auth / Home
Dashboard
Bot

💡 Inspiration

The project was inspired by the struggles that professionals and students face when trying to study through lengthy YouTube videos and playlists. Browsing through countless hours of content to find specific topics of interest can be frustrating and time-consuming. The goal was to create a solution that serves as a helpful companion, making the study process more efficient and less cumbersome.

🎯 What it does

The "YouTube Study Companion" is designed to simplify the process of navigating long YouTube videos and playlists. Users provide a video or playlist link, and the bot identifies and returns timestamps where specific topics of interest are discussed based on user queries. Instead of manually skimming through content, users can instantly jump to relevant sections, making the learning experience more streamlined.

🛠️ How we built it

We used a combination of technologies to bring this idea to life. The backend utilizes TiDB for vector similarity search and MySQL for database management. Jina AI’s embedding API plays a key role in enabling the bot’s intelligence, while NLP techniques through NLTK are used for stopword removal and further query refinement. The frontend is built as a web app using React with Next.js and Django, providing an intuitive and user-friendly interface.

🧗‍♂️ Challenges we ran into

Building a system that accurately identifies relevant timestamps in lengthy videos was a technical challenge, especially in balancing accuracy and speed. Implementing similarity search efficiently while managing large data sets was also a hurdle, along with making sure the system scales well for different user queries and content lengths.

🏆 Accomplishments that we're proud of

We're proud of creating an efficient tool that makes a real impact on learners who rely on YouTube content for their studies. Our use of generative AI in a novel, agentic flow, rather than a simple retrieval-augmented generation (RAG) approach, sets our solution apart. Additionally, the integration of TiDB Cloud for efficient vector search and handling large-scale queries was a key accomplishment.

📚 What we learned

Through this project, we gained deeper insights into the power of similarity search and generative AI. We learned to leverage advanced technologies like TiDB Cloud in innovative ways to make search processes more efficient. Furthermore, working with AI-driven tools and combining them with web technologies like React and Django gave us a more holistic understanding of end-to-end application development.

🚀 What's next

In the future, we aim to extend the functionality of the YouTube Study Companion by adding features like note-taking and summarization, making it a more comprehensive study tool. We also plan to expand the data sources it supports and eventually port it into a browser extension for even more seamless integration into users' workflows.

📎Source Code

📎API Documentation

📎Presentation