Self-folding enables compact, deployable and reconfigurable structures for close and far remote applications. As the degree-of-freedom in achieving folded shapes is dictated by the number of actuators, distributed on a quasi-2-D lattice, their design remains challenging in terms of size, weight, efficiency, repeatability and fabrication. While the traditional electric motors are difficult to downscale and assemble, the active material-based actuators consume high power and are limited by slow and irreversible motions. Here we present a new selective fluid-driven actuation and embedment method for constructing multi-creased, self-folding and reversible robotic origami structures. Our design enables an underactuated, yet a distributed control of origami hinges with single and multi-step folding sequence by programming the fluidic networks. The proposed tunable fluidic actuator-channel networks are compactly embedded in a rapid multi-layer lamination process with minimal assembly. We provide an analytic model for the folding actuator unit, validate it with several prototypes and demonstrate its integration into complex networks for folding multi-hinged tessellations, including Miura-ori and sequentially folding box patterns.