BACKGROUND AND HYPOTHESIS: The majority of first episode psychosis (FEP) patients are undetected (DET-) by clinical high-risk for psychosis (CHR-P) services prior to onset and therefore do not receive preventive care for psychosis. We compared features of the psychosis prodrome (symptoms and substance use) between DET- and FEP patients detected by CHR-P services (DET+) to determine whether they share a common prodromal phase. STUDY DESIGN: Retrospective Reporting of Studies Conducted Using Observational Routinely Collected Health Data statement-compliant electronic health record cohort study. We extracted 65 prodromal features (symptoms and substance use before FEP onset) using natural language processing to assess the presence, duration, and first presentation of the psychosis prodrome and occurrences of features across the prodrome. Duration and feature occurrences were compared between DET+ and DET- individuals using Mann-Whitney U tests and Wilcoxon Effect Size, while presence and first presentation were compared using logistic regression. STUDY RESULTS: A total of 1545 FEP patients (n = 119 DET+ [mean age 28.7 years; SD = 9.4; 61.6% male]) were included. There were no significant differences in the presence (DET + =85.0%, DET- = 85.6%, P = .83) or duration (DET + =18.8 months, DET- = 18.4 months, P = .89) of the psychosis prodrome. There were no significant differences in first presentation of psychotic symptoms between groups (Pcorr > .05). Frequency of occurrences of thought broadcasting (r = 0.07, Pcorr = .04) was higher and hostility (r = 0.08, Pcorr = .04) lower in DET+ compared to DET- across the prodrome, though effect sizes were small. CONCLUSIONS: DET+ and DET- individuals experience similar psychosis prodromes prior to FEP onset. DET- individuals can likely be identified earlier if detection strategies are improved.