Microfluidics is an evolving scientific field with immense commercial potential: analytical applications, such as biochemical assay development, biochemical analysis and biosensors as well as chemical synthesis applications essentially require microfluidics for sample handling, treatment or readout. A number of techniques are available to create microfluidic structures today. On industrial scale replication techniques such as injection molding are the gold standard whereas academic research mostly focuses on replication by casting of soft elastomers such as polydimethylsiloxane (PDMS). Both of these techniques require the creation of a replication master thus creating the microfluidic structure only in the second process step—they can therefore be termed two-(or multi-)step manufacturing techniques. However, very often the number of pieces to be created of one specific microfluidic design is low, sometimes even as low as one. This raises the question if two-step manufacturing is an appropriate choice, particularly if short concept-to-chip times are required. In this case one-step manufacturing techniques that allow the direct creation of microfluidic structures from digital three-dimensional models are preferable. For these processes the number of parts per design is low (sometimes as low as one), but quick adaptation is possible by simply changing digital data. Suitable techniques include, among others, maskless or mask based stereolithography, fused deposition molding and 3D printing. This work intends to discuss the potential and application examples of such processes with a detailed view on applicable materials. It will also point out the advantages and the disadvantages of the respective technique. Furthermore this paper also includes a discussion about non-conventional manufacturing equipment and community projects that can be used in the production of microfluidic devices.