Microfluidic devices are used extensively in the development of new in vitro cell culture models like organs-on-chips. A typical feature of such devices is the patterning of biological hydrogels to offer cultured cells and tissues a controlled three-dimensional microenvironment. A key challenge of hydrogel patterning is ensuring geometrical confinement of the gel, which is generally solved by inclusion of micropillars or phaseguides in the channels. Both of these methods often require costly cleanroom fabrication, which needs to be repeated even when only small changes need be made to the gel geometry, and inadvertently expose cultured cells to non-physiological and mechanically stiff structures. Here, we present a technique for facile patterning of hydrogel geometries in microfluidic chips, but without the need for any confining geometry built into the channel. Core to the technique is the use of laminar flow patterning to create a hydrophilic path through an otherwise hydrophobic microfluidic channel. When a liquid hydrogel is injected into the hydrophilic region, it is confined to this path by the surrounding hydrophobic regions. The various surface patterns that are enabled by laminar flow patterning can thereby be rendered into three-dimensional hydrogel structures. We demonstrate that the technique can be used in many different channel geometries while still giving the user control of key geometric parameters of the final hydrogel. Moreover, we show that human umbilical vein endothelial cells can be cultured for multiple days inside the devices with the patterned hydrogels and that they can be stimulated to migrate into the gel under the influence of trans-gel flows. Finally, we demonstrate that the patterned gels can withstand trans-gel flow velocities in excess of physiological interstitial flow velocities without rupturing or detaching. This novel hydrogel-patterning technique addresses fundamental challenges of existing methods for hydrogel patterning inside microfluidic chips, and can therefore be applied to improve design time and the physiological realism of microfluidic cell culture assays and organs-on-chips.