A brain tumor is defined as an excessive increase in neurons in the brain. Benign and malignant are the two main categories of brain tumors, which are defined by an aberrant growth of neurons in the brain. Malignant brain tumors can spread rapidly to other parts of the brain and spine, causing serious health dangers, whereas benign brain tumors develop slowly, are rare, and do not metastasis. Determining the best treatment interventions and enhancing patient outcomes depend on the timely and precise categorization of brain tumors. In this work, a novel artificial intelligence-based paradigm for classifying brain tumors from magnetic resonance imaging (MRI) data is proposed. We combine cuttingedge machine learning algorithms, such as eXtreme Gradient Boosting (XGB) and Light Gradient Boosting Machine (LGBM), with cuttingedge feature extraction techniques like Grey Level Co-occurrence Matrix (GLCM), Local Binary Patterns (LBP), and Gabor filters to achieve a high classification accuracy of 98.53%, with a precision of 98.47% and an F1-score of 98.44 %. We use feature importance techniques like SHAP (SHapley Additive exPlanations) and Eli5 to further improve the interpretability of our model by revealing which features have the most influence on the model's conclusions. For clinical application, this degree of openness is crucial because it helps healthcare providers comprehend and have more faith in the insights our system produces. The proposed framework demonstrates strong potential as a non-invasive diagnostic aid for early and accurate tumor classification, aiding in more effective clinical decision-making.