This paper presents a new lossless, cluster-based randomly-accessible and progressive 3D triangular mesh compression scheme. The scheme is targeted at the visualization of large meshes, supporting the encoding of vertex colors. We have extended the previous approach of [3] by using the progressive mesh encoder of [6] to compress the face clusters. The schemes were modified to prevent duplication of the boundary geometry information between cluster vertices. We propose a visualization framework that exploits the features of our codec, with solutions given to solve the problem of cracks appearing between clusters during visualization. Results demonstrate that acceptable compression ratios can be achieved using the introduced features.