Recommender systems are essential for guiding users to relevant items and locations; however, the cold-start problem caused by missing or unrated items remains a persistent challenge. This study proposes a novel hybrid framework that integrates the item-based clustering hybrid method (ICHM) with the Slope One algorithm to specifically address cold-start scenarios in traditional craft recommender systems. A unique dataset of 48 craft locations and 60 traditional Balinese craft products, collected through direct field observation, representing an original contribution that bridges cultural heritage and advanced recommendation technologies, was used for validation. The framework predicts missing ratings using Slope One and generates recommendation scores via a weighted-sum function, providing dual recommendations for both products and production locations. The experimental results indicate high prediction accuracy, with overall mean absolute error values well below acceptable thresholds, confirming the system’s reliability and robustness. Beyond technical contributions, it highlights the socio-economic and cultural potential of RSs in preserving and promoting local heritage. © 2026 Faculty of Engineering, Universitas Indonesia. All rights reserved.