Molecular recognition is of paramount importance for modern chemical processes and has now been achieved for small molecules using well-established host-guest chemistry and adsorption-science principles. In contrast, technologies for recognizing polymer structure are relatively undeveloped. Conventional polymer separation methods, which are mostly limited in practice to size-exclusion chromatography and reprecipitation, find it difficult to recognize minute structural differences in polymer structures as such small structural alterations barely influence the polymer characteristics, including molecular size, polarity, and solubility. Therefore, most of the polymeric products being used today contain mixtures of polymers with different structures as it is challenging to completely control polymer structures during synthesis even with state-of-the-art substitution and polymerization techniques. In this context, development of novel techniques that can resolve the challenges of polymer recognition and separation is in great demand, as these techniques hold the promise of a new paradigm in polymer synthesis, impacting not only materials chemistry but also analytical and biological chemistry.In biological systems, precise recognition and translation of base monomer sequences of mRNA are achieved by threading them through small ribosome tunnels. This principle of introducing polymers into nanosized channels can possibly help us design powerful polymer recognition and separation technologies using metal-organic frameworks (MOFs) as ideal and highly designable recognition media. MOFs are porous materials comprising organic ligands and metal ions and have been extensively studied as porous beds for gas separation and storage. Recently, we found that MOFs can accommodate large polymeric chains in their nanopores. Polymer chains can spontaneously infiltrate MOFs from neat molten and solution phases by threading their terminals into MOF nanochannels. Polymer structures can be recognized and differentiated due to such insertion processes, resulting in the selective adsorption of polymers on MOFs. This enables the precise recognition of the polymer terminus structure, resulting in the perfect separation of a variety of terminal-functionalized polymers that are otherwise difficult to separate by conventional polymer separation methods. Furthermore, the MOFs can recognize polymer shapes, thus enabling the large-scale separation of high purity cyclic polymers from the complex crude mixtures of linear polymers, which are used as precursor materials in common cyclization reactions. In solution-phase adsorption, many factors, including molecular weight, terminal groups, polymer shape, polymer-MOF interaction, and coexisting solvent molecules, influence the selective adsorption behavior; this yields a new liquid chromatography-based polymer separation technology using an MOF as the stationary phase. MOF-packed columns, in which a novel separation mode based on polymer insertion into the MOF operates under a dynamic insertion/rejection equilibrium at the liquid/solid interface, exhibited excellent polymer separation capability. The polymer recognition principle described in this study thus has a high probability for realizing previously unfeasible polymer separations based on monomer composition and sequences, stereoregularity, regioregularity, helicity, and block sequences in synthetic polymers and biomacromolecules.