A prevalent technique in finite element modeling for emulating fractures involves the use of cohesive models with zero-thickness interface elements, enabling the discrete representation of cracks [1]. One significant limitation of this method is the increased computational load, primarily due to the duplication of nodes when interface elements are integrated and the tendency of cracks to follow a mesh-dependent path along the boundaries of elements. To mitigate the computational burden, one strategy is the adaptive insertion of interface elements during analysis, effectively reducing the number of node duplications and, consequently, computational demands [2]. However, the persistent challenge of mesh-dependent crack paths still requires attention. Various strategies have been proposed to diminish this bias in the crack paths of two-dimensional finite element models. These methods involve either repositioning or splitting elements to align their edges in directions determined by stress or energy criteria. Nonetheless, there is a scarcity of solutions for three-dimensional models [3]. This contribution introduces a comprehensive computational approach for forecasting crack initiation and progression in both two- and three-dimensional finite element models. This is achieved through the adaptive insertion of interface elements, which are augmented with a cohesive law and a crack path prediction mechanism based on local tractions of interface elements. The mesh is then adaptively adjusted by reorienting the elements such that their boundaries align with the computed crack path. The efficiency of this novel method is validated through classic benchmark tests, including the L-shaped panel and the Nooru-Mohammed test. [1] Gudžulić V., Neu, G.E., Gebuhr, G. Anders, S. and Meschke, G., 'Numerical multi-level model for fiber-reinforced concrete – Multi-level validation based on an experimental study on high-strength concrete' (in German), Beton- und Stahlbetonbau (2019), Ernst & Sohn. [2] Paulino, G. H., Celes, W., Espinha, R., Zhang, Z., A general topology-based framework for adaptive insertion of cohesive elements in finite element meshes, Eng. Comput., vol. 24, pp. 59-78, 2008. [3] Kaczmarczyk, Ł., Nezhad, M. M. and Pearce, C., Three-dimensional brittle fracture: configurational-force-driven crack propagation, Int. J. Numer. Methods Eng., vol. 97, no. 7, pp. 531–550, 2014.