COBALT is a constraint-based alignment tool designed to align multiple protein sequences. It combines pairwise constraints derived from database searches, sequence similarity, and user input to create a progressive multiple alignment. The tool leverages the conserved domain database (CDD) and PROSITE protein-motif database to improve alignment quality. COBALT's performance is comparable to or better than other popular alignment tools like ClustalW, MUSCLE, ProbCons, and PCMA, while maintaining reasonable runtime. The algorithm uses a general framework that includes steps for finding constraints, generating a guide tree, computing the multiple sequence alignment, and refining the alignment. COBALT is available in the NCBI C++ toolkit and has been tested on various benchmark datasets, demonstrating its effectiveness in multiple alignment tasks.COBALT is a constraint-based alignment tool designed to align multiple protein sequences. It combines pairwise constraints derived from database searches, sequence similarity, and user input to create a progressive multiple alignment. The tool leverages the conserved domain database (CDD) and PROSITE protein-motif database to improve alignment quality. COBALT's performance is comparable to or better than other popular alignment tools like ClustalW, MUSCLE, ProbCons, and PCMA, while maintaining reasonable runtime. The algorithm uses a general framework that includes steps for finding constraints, generating a guide tree, computing the multiple sequence alignment, and refining the alignment. COBALT is available in the NCBI C++ toolkit and has been tested on various benchmark datasets, demonstrating its effectiveness in multiple alignment tasks.