Background In high-income countries, early intervention in psychosis (EIP) services provide treatment for first episode psychosis (FEP) as early as possible, proving to effectively improve health outcomes. These services rarely exist in low- and lower-middle-income countries (LMICs). The TRANSLATE programme aims to evaluate the effectiveness of EIP services in terms of implementation outcomes in two LMIC countries, and to develop a clinical tool to predict the risk of Treatment Resistant Schizophrenia (TRS) in patients with FEP. Methods This is an observational implementation research study, conducted across five sites in Pakistan and one in Sri Lanka, with recruitment of 670 FEP patients. The key clinical such as status of functioning, quality of life, psychopathology, family burden, stigma and physical health, will be assessed at baseline and periodically over 1 year period. A descriptive analysis of baseline data and service utilisation data will be conducted, stratified by site. For implementation outcomes, the probabilities of disengagement and remission will be modelled using logistic regression models adjusting for duration of untreated psychosis, baseline symptom severity, and geographical location. For secondary implementation outcomes, selected key clinical outcomes and a process evaluation will be conducted. To develop a clinical model prediction model for TRS, a penalised multivariable logistic regression model will be fitted to estimate the one-year risk of TRS following an FEP diagnosis, incorporating up to 15 predictor variables identified within the study cohort. Model performance will be assessed using calibration plots and the C-statistics for discrimination. Internal validation will be carried out by bootstrapping, and the optimism-adjusted calibration slope will be used as a global shrinkage factor to account for overfitting, with the model intercept re-estimated to ensure calibration-in-the-large. Discussion This will be the first large study that will implement and evaluate the EIP service in a LMIC country setting. The findings of this study and the process evaluation will inform the adoption and scaling-up of EIP services in routine clinical practice. The clinical prediction model for the early detection of TRS will help initiate effective treatments for TRS at the earliest possible stage and potentially reduce the significant burden of disease associated with TRS.