The majority of viruses within the human gut are obligate bacterial viruses known as bacteriophages (phages in short)1. As strict bacterial viruses, the study of phages in the gut has always taken a bacteriocentric view. Gut phages are known to sustain top-down ecological control2-4 and co-evolve with bacterial communities5. This has traditionally been investigated via empirical in vitro approaches, and more recently through the use of in vivo animal models6-8, which allows us to probe phage-bacteria populations within the gut4,9. Nevertheless, studies have fixated on phage-bacteria co-evolution while ignoring potential phage-mammalian interactions. Here, we investigated the evolution between the phage and the mammalian “host”. We established in vitro lab-on-a-chip devices (a.k.a. gut-on-a-chip) to recapitulate a life-like gut mucosa, which we inoculated with lytic T4 phage and its Escherichia coli bacterial host. By sampling the gut-on-a-chip at high temporal resolution, we showed that the mucosal environment supports stable phage-bacteria co-existence. Experimental evolution followed by deep sequencing of phage populations revealed that phages were directly adapting to the mucosa through de novo mutations. We demonstrated that high phage multiplicities-of-infection sustained within the mucosa promoted genetic recombination as a mechanism of adaptation. Importantly, a single mutation in the phage capsid protein, Hoc – known to facilitate phage adherence to mucus10 – resulted in an altered mucus glycan-binding phenotype. Through a competition assay, we showed that mucus-evolved phage outcompeted the ancestral wildtype phage within the gut-on-a-chip. Collectively, our study revealed that phages – despite their long-held bacteriotropism – are able to evolve in direct response to the mammalian host.