This study aims to develop a dependable Cafe Recommender System specifically tailored for the Bandung area by employing a fusion of Item-Based Collaborative Filtering (IBCF) and Recurrent Neural Network (RNN) methodologies. The motivation stems from the need for more accurate and relevant café recommendations in Bandung, a city known for its diverse cafe scene. Previous research often relied on either collaborative filtering or natural language processing approaches independently, leading to limitations in capturing the nuances of user preferences and sentiments. To address this, we leverage IBCF to analyze user rating data and identify similarities amongst cafes, generating personalized recommendations. Concurrently, we employ RNN to examine and understand user reviews, facilitating a more contextually sensitive suggestion procedure. We hypothesize that this hybrid approach will enhance recommendation precision and pertinence. The model was evaluated using a dataset of 6854 rows of user ratings and reviews. The assessment, conducted using Precision, Recall, and F1-score, yielded promising results: 89.04%, 88.75%, and 88.62%, respectively. Based on our experiments, the model consistently recommended cafes such as are Little Contrast Riau, Little Contrast Braga, Northwood Fast Casual Dining Cihampelas, Mampirry Cafe & Resto Kopo, and Sama Dengan Cipaganti for users seeking a unique and satisfying cafe experience in Bandung. These cafes were frequently associated with positive sentiments in user reviews and aligned well with the preferences inferred from user rating patterns, showcasing the effectiveness of the hybrid IBCF-RNN approach in providing tailored cafe recommendations.