-
-
Landing Page-1st time
-
Landing Page-Start loading
-
Analyzing
-
Report Page
Inspiration
YouTube has become one of the most essential platforms for accessing a wide range of information. However, when it comes to specific topics like finance, travel, or technology, users often face the challenge of sifting through a large volume of videos to find the most relevant and valuable content. This is especially true for popular or recommended videos, where the sheer number of options can feel overwhelming. Our inspiration for this project came from the need to help users save time and effort. By providing a tool that extracts the key takeaways from videos, we aim to make YouTube a more efficient platform for learning and discovering useful information, allowing users to quickly grasp the essential points without having to watch entire videos.
What it does
We developed a YouTube video summarization tool that helps users quickly extract and comprehend the key points of videos. For instance, "stock trading tutorials" or "traveling in the U.S." The tool delivers concise summaries of videos, ensuring that users don’t miss out on valuable insights. It allows them to efficiently grasp the essence of the content without needing to watch the entire video.
How we built it
We developed a React-based web application with a Node.js backend serving as API endpoints for external API integration. On the front end, we incorporated the YouTube Data API V3 and a video transcription API as key data sources. However, the standout feature is our AI-driven functionality. Using the Summarize API, we generate concise video overviews in "Headline" mode and detailed key points in "Key Points" mode, giving users a clear and summarized context of the video content. But we didn’t stop there. To add even more value, we integrated the Prompt API with carefully designed prompts to generate a structured, unordered list of action items related to the video. This ensures users not only understand the video's content but also know what questions to ask or actions to take based on it. To make our product accessible to a broader audience, we deployed the front-end application on GitHub Pages, while hosting the backend on Render.
Challenges we ran into
Initially, we used the Summarize API to process information, but we struggled to generate actionable items or a clear to-do list for users, which limited our ability to provide practical next steps directly. Our goal was to refine our approach to enable the generation of clear action items that users could easily follow. To address this, we moved beyond relying solely on the Summarize API. We used it to create an overview and then integrated the Prompt API to derive actionable items. During this process, we focused heavily on refining the wording, prioritizing precision and readability. This ensured that the final content generated via the Prompt API was user-friendly and easy to understand. By combining the two APIs and iterating on the clarity of our language, we successfully resolved the initial challenge. The solution provided both a high-level overview and actionable steps, significantly improving the user experience.
Accomplishments that we're proud of
We successfully delivered a more user-friendly summary output by blending summarization and prompt generation. This combination allowed us to create results that are richer, easier to understand, and better aligned with our initial user cases, ensuring that the tool meets the needs of users effectively.
What we learned
Through this project, we learned that combining multiple APIs can unlock new possibilities for solving complex problems, making solutions more versatile and better suited to real-world needs. Each API specializes in a particular function, and coordinating their outputs felt like leading a team of experts. By orchestrating their unique functionalities and leveraging them in a coordinated manner, we were able to uncover new possibilities for solving complex problems.
What's next for Shio AI
Looking ahead, we plan to further enhance the efficiency of information processing, helping users absorb content even faster and more effectively. Building on this foundation, we aim to expand our tool’s capabilities into the education industry, creating an even more powerful learning resource for the modern age, where short videos and rapid content consumption are the norm. Our goal is to provide users with an innovative learning experience that keeps up with the fast-paced world of digital content.
Log in or sign up for Devpost to join the conversation.