Endosymbiotic bacteria associated with eukaryotic hosts are omnipresent in nature, particularly in insects. Studying the bacterial side of host-symbiont interactions is, however, often limited by the unculturability and genetic intractability of the symbionts. <i>Spiroplasma poulsonii</i> is a maternally transmitted bacterial endosymbiont that is naturally associated with several <i>Drosophila</i> species. <i>S. poulsonii</i> strongly affects its host's physiology, for example by causing male killing or by protecting it against various parasites. Despite intense work on this model since the 1950s, attempts to cultivate endosymbiotic <i>Spiroplasma in vitro</i> have failed so far. Here, we developed a method to sustain the <i>in vitro</i> culture of <i>S. poulsonii</i> by optimizing a commercially accessible medium. We also provide a complete genome assembly, including the first sequence of a natural plasmid of an endosymbiotic <i>Spiroplasma</i> species. Last, by comparing the transcriptome of the <i>in vitro</i> culture to the transcriptome of bacteria extracted from the host, we identified genes putatively involved in host-symbiont interactions. This work provides new opportunities to study the physiology of endosymbiotic <i>Spiroplasma</i> and paves the way to dissect insect-endosymbiont interactions with two genetically tractable partners. <b>IMPORTANCE</b> The discovery of insect bacterial endosymbionts (maternally transmitted bacteria) has revolutionized the study of insects, suggesting novel strategies for their control. Most endosymbionts are strongly dependent on their host to survive, making them uncultivable in artificial systems and genetically intractable. <i>Spiroplasma poulsonii</i> is an endosymbiont of <i>Drosophila</i> that affects host metabolism, reproduction, and defense against parasites. By providing the first reliable culture medium that allows a long-lasting <i>in vitro</i> culture of <i>Spiroplasma</i> and by elucidating its complete genome, this work lays the foundation for the development of genetic engineering tools to dissect endosymbiosis with two partners amenable to molecular study. Furthermore, the optimization method that we describe can be used on other yet uncultivable symbionts, opening new technical opportunities in the field of host-microbes interactions.