What if the scoreboard had no delay?

A team of developers has introduced an open-source solution to eliminate delays in live esports scoreboards, specifically for games like BGMI that lack public APIs. By utilizing a local OBS-based overlay, the tool monitors stream footage in real-time to detect kill feed events. The system employs a sophisticated filtering pipeline that reduces the processing load by 99.3%, allowing it to handle 60 FPS streams efficiently. Using a combination of PaddleOCR and a fine-tuned YOLOv11 model, the application identifies player actions and calculates points instantly, providing viewers with an immediate and accurate leaderboard. This project, developed for the Hacktoberfest Open-Source AI Challenge, demonstrates how multimodal machine learning can bridge the gap in game data accessibility, offering a lightweight and effective alternative for tournament organizers who previously struggled with manual compilation of points tables during live broadcasts.