This paper introduces a novel approach for detecting surgical instruments in endoscopic videos using the YOLOv8 model enhanced with Oriented Bounding Boxes (OBBs). By incorporating orientation and rotation information, this method overcomes the limitations of traditional axis-aligned bounding boxes, providing more accurate detection and tracking of surgical tools during minimally invasive surgeries (MIS). A custom dataset, m2cai16-tool-obb, was developed and used to train and evaluate the model. The YOLOv8x-OBB model demonstrated high performance, achieving a precision of 0.802, recall of 0.609, and mAP50 of 0.703. Additionally, we developed a web interface based on the Streamlit library, enabling real-time surgical tool detection and visualization during MIS procedures. The proposed system has the potential to significantly improve the accuracy and efficiency of surgical tool tracking in endoscopic environments.