Shrenika Ghimire | Hello and Namaste, Everyone! I am Shrenika Ghimire, a student at The Westminster College, and I’m excited to share one of the most meaningful projects I’ve worked on SportsAIght. This was our hackathon submission, built to solve a real problem faced by athletes in countries like Nepal: not being seen. SportsAIght is an AI-powered system that automatically creates sports highlights with real-time commentary and captions. We aimed to make these tools accessible, emotional, and fast, especially for those who lack access to major media resources.

In today’s sports world, professional players are constantly in the spotlight. Their matches are streamed live, cut into highlights within seconds, and shared globally. But that’s not the case for everyone. In places like Nepal, many skilled athletes play amazing matches, but their moments go undocumented. With limited access to equipment, editing tools, or technical skills, these athletes don’t get the recognition they deserve. As one local player told us, “Even when I score a stunning goal, no one sees it. It feels like my effort disappears after the whistle.”

That quote became the core of our project. We realised this was not just a tech challenge it was a visibility issue. Our mission with SportsAIght was to build a solution that makes every player feel seen, no matter their background or budget.

So, how does SportsAIght work? It starts by analysing a sports video using motion-based segmentation, which helps detect the moments when something exciting is happening. Then, it uses YOLOv8, a powerful object detection model, to identify specific events like football kicks, player movement, or crowd surges. These key moments are automatically clipped using FFmpeg, meaning there’s no need to cut videos manually.

But we wanted more than just highlight clips. We wanted to tell the story of each moment. That is where GPT-3.5 Turbo comes in. This AI model generates real-time commentary and captions, turning each highlight into a short, shareable, and emotional experience just like you would see on ESPN or YouTube Shorts.
We built the backend in Python 3 and the frontend using TypeScript. The final output is a fully captioned MP4 video, perfect for athletes to share on platforms.

SportsAIght is not just for football players. It can be used by sports newsrooms, coaches for performance reviews, and even broadcasting platforms. It’s especially valuable for developing nations where advanced video tools are unaffordable. Our goal was to create something light, scalable, and affordable so everyone, regardless of resources, could access it.

In the future, we plan to support other sports like cricket and basketball, build a live web interface, and deploy the system on real platforms. This would allow schools, colleges, and clubs in Nepal to highlight their players in real time.

Building SportsAIght taught me that technology is not just about innovation, it is about impact. With the right tools, we can uplift communities and bring untold stories into the spotlight. I’m proud of what our team created, and I look forward to where this idea can go next.