Hundreds of messenger RNAs (mRNAs) are transported into neurites to provide templates for the assembly of local protein networks. These networks enable a neuron to configure different cellular domains for specialized functions. According to current evidence, mRNAs are mostly transported in rather small packages of one to three copies, rarely containing different transcripts. This opens up fascinating logistic problems: how are hundreds of different mRNA cargoes sorted into distinct packages and how are they coupled to and released from motor proteins to produce the observed mRNA distributions? Are all mRNAs transported by the same transport machinery, or are there different adaptors or motors for different transcripts or classes of mRNAs? A variety of often indirect evidence exists for the involvement of proteins in mRNA localization, but relatively little is known about the essential activities required for the actual transport process. Here, we summarize the different types of available evidence for interactions that connect mammalian mRNAs to motor proteins to highlight at which point further research is needed to uncover critical missing links. We further argue that a combination of discovery approaches reporting direct interactions, in vitro reconstitution, and fast perturbations in cells is an ideal future strategy to unravel essential interactions and specific functions of proteins in mRNA transport processes.