Micro-expressions have the potential to reveal true emotions in fields such as public safety and healthcare. However, due to their small movement and short duration, their spotting has always been a significant challenge. To address the limitations of existing methods in feature extraction and representation, this paper proposes a micro-expression spotting method that combines Optical Flow features with Categorical Text information (OFCT). First, facial alignment is performed, and optical flow features in both the main and secondary directions are extracted from 13 regions of interest to capture subtle movements more comprehensively. Then, the micro-expression category text information is vectorized and fused with the main and secondary direction optical flow features, enhancing the spotting ability at the semantic level. In the micro-expression segment spotting phase, low-pass filtering and empirical mode decomposition are used to smooth and denoise the optical flow curves, preserving key change trends. Finally, high-precision micro-expression segments are obtained by combining threshold determination with non-maximum suppression. Experimental results show that the proposed method outperforms existing mainstream methods on the CAS(ME)² and SAMM-LV databases. Additionally, its F1-score on the MEGC2022 test set also exceeds the best result at the time, demonstrating the effectiveness and robustness of integrating category text information with main and secondary direction optical flow features in micro-expression spotting.