BACKGROUND: Mental health professionals face unique demands and stressors in their work, resulting in high rates of burnout and distress. Clinical supervision is a widely adopted and valued mechanism of professional support, development, and accountability, despite the very limited evidence of specific impacts on therapist or client outcomes. The current study aims to address this by exploring how psychotherapists develop competence through clinical supervision and what impact this has on the supervisees' practice and their clients' outcomes. This paper provides a rationale for the study and describes the protocol for an in-depth qualitative study of supervisory dyads, highlighting how it addresses gaps in the literature. METHODS/DESIGN: The study of 16-20 supervisor-supervisee dyads uses a qualitative mixed method design, with two phases. In phase one, supervisors who are nominated as expert by their peers are interviewed about their supervision practice. In phase two, supervisors record a supervision session with a consenting supervisee; interpersonal process recall interviews are conducted separately with supervisor and supervisee to reflect in depth on the teaching and learning processes occurring. All interviews will be transcribed, coded and analysed to identify the processes that build competence, using a modified form of Consensual Qualitative Research (CQR) strategies. Using a theory-building case study method, data from both phases of the study will be integrated to develop a model describing the processes that build competence and support wellbeing in practising psychotherapists, reflecting the accumulated wisdom of the expert supervisors. DISCUSSION: The study addresses past study limitations by examining expert supervisors and their supervisory interactions, by reflecting on actual supervision sessions, and by using dyadic analysis of the supervisory pairs. The study findings will inform the development of future supervision training and practice and identify fruitful avenues for future research.