Capacitive deionization (CDI) technology has attracted a great deal of attention over the past few decades due to its high energy efficiency (less than 1 kWh/m-3) compared to conventionally utilized reverse osmosis (2–4 kWh/m-3) and distillation based (50–80 kWh/m-3) desalination technologies. However, due to a couple of issues including cost and requirement for a disruptive discharging step, it has been regarded as a challenging task to perform large-scale desalination of seawater by the CDI process. Herein, we report a novel design for a 3D desalination system utilizing porous lattice scaffolds. It can desalinate water as salty as sea water (35 g/L) with desalting efficiency comparable with conventional flow-electrode-based CDI (FCDI). By coating of ion exchange membranes and a graphene layer inside each channel of the lattice structures, honeycomb-shaped desalination cells could be realized. First, the desalination performance of the new design was investigated in batch mode. It exhibited comparable desalting performance while maintaining typical advantages of FCDI systems, such as low energy consumption and no need of a discharging step. Furthermore, since the porous structures involve channels for ion transportation and act as a structural scaffold, the cell architecture is remarkably compact and it can be readily scaled-up by varying the number of unit cells and its configuration, allowing great increase in the salt removal capacity. We scaled-up the desalination system to a 3 × 3 cell, which showed desalination efficiency four times higher than that of a 1 × 3 cell. Such facile build-up of the unit cell structure and versatile design of cell configuration provide strong potential for further scalability. We also determined that this system can be operated successfully in continuous mode, indicating that neither a distinct discharging step nor substitution of inflow and effluent are required. Because there is opportunity for much more improvement of desalting performance in terms of the electrode materials, resistance at the interface, and conductivity of the current collecting layer of this system, we expect that our approach has opened a new door in the field of desalination research.