Background: Moral conservatism within government and communities has resulted in a reluctance to support the provision of opioid agonist therapy for people with opioid use disorders in South Africa. In April 2017, South Africa's first low-threshold opioid agonist therapy demonstration project was launched in Durban. The project provided 54 low-income people with heroin use disorders methadone and voluntary access to psychosocial services for 18 months. At 12 months, retention was 74%, notably higher than the global average. In this paper, we aim to make sense of this outcome. Methods: Thirty semi-structured interviews, two focus groups, ten oral histories and ethnographic observations were done at various project time points. These activities explored participants' pathways into drug use and the project, their meaning attributed to methadone, the factors contributing to project success and changes they experienced. Recordings, transcripts, notes and feedback were reviewed and triangulated. Key factors contributing to retention were identified and analysed in light of the existing literature. Results: The philosophy and architecture of the project, and social cohesion were identified as the main factors contributing to retention. The use of a harm reduction approach enabled participants to set and be supported to achieve their treatment goals, and was shown to be important for the development of trusting therapeutic relationships. The employment of a restorative justice paradigm provided a sense of acceptance of humanity and flaws as well as an imperative to act responsibly towards others, fostering a culture of respect. Social cohesion was fostered through the facilitation of group sessions, a peace committee and group sport (soccer). In concert, these activities provided opportunities for participants to demonstrate care and interest in one another's life, leading to interdependence and care, contributing to them remaining in the project. Conclusions: We believe that the high retention was achieved through attraction. We argue that opioid agonist therapy programmes should take the principles of harm reduction and restorative justice into consideration when designing low-threshold opioid agonist therapy services. Additionally, ways to support cohesion amongst people receiving agonist therapy should be explored to support their effective scale-up, both in low-middle income countries and in high-income countries.