Objectives To evaluate the diagnostic accuracy of the Roche SARS-CoV-2 & Flu A/B Rapid Antigen Test at the point of care. Design Prospective diagnostic accuracy study. Setting 17 primary care practices in England. Participants 500 individuals with symptoms consistent with possible SARS-CoV-2 or influenza infection identified upon presentation to primary care or via medical note review. Primary and secondary outcome measures Sensitivity, specificity and predictive values, compared to a laboratory reference standard of real-time reverse transcription PCR, using samples collected using a combined nasal and oropharyngeal swab. Results Of 481 participants with available index and reference test results, 5.6% (27/481) were reference standard-positive for SARS-CoV-2, 11.4% (55/481) for Influenza A and 1.9% (9/481) for Influenza B. The sensitivity of the antigen test to detect SARS-CoV-2 was 70.4% (19/27, 95% CI 49.6–86.2%) and specificity was 99.3% (451/454, 95%CI 98.1–99.9%). For Influenza A, sensitivity was 29.1% (16/55, 95% CI 17.6–42.9%) and specificity 98.6% (420/426, 97.0–99.5%), and for Influenza B, sensitivity was 22.2% (2/9, 2.8–60.0%) and specificity 98.1% (463/472, 96.4–99.1%). Conclusions In a primary care population of symptomatic individuals, the assay was highly specific and had moderate sensitivity to detect SARS-CoV-2, but did not detect the majority of influenza infections. Registration ISRCTN14226970.