Objective: Low back pain (LBP), a common public health problem affecting nearly 80 % of the population, causes disability and work absence, while exercise is recommended as an adjuvant treatment. YouTube is one of the most frequently used source of information, especially by patients. The aim of the study was to analyze clinically the reliability, feasibility and the content quality of exercises videos recommended for LBP on YouTube. channel. Methods: Based on the observational study of the number of views of YouTube videos, 154 videos worth evaluation were selected using the keywords "low back pain" and "exercises",. The content quality of each video was evaluated with a) the American Medical Association Journal (JAMA), b) DISCERN, c) Global Quality Score (GQS), and d) Medical Quality Video Assessment Tool (MQ-VET) scales, while Pearson correlation analysis was performed to determine the correlation between the JAMA, GQS, DISCERN, MQ-VET scale scores. Results: According to the findings, the primary source for uploading exercise videos were physicians (32.5 %), with a high content quality. There was a statistically significant positive correlation between the total scores of all quality scales and the independent variables defined as view rates and number of comments (p < .05). The analysis of the video content showed that only 35 % (n = 54) of the videos explained the risks of the recommended exercises in detail Conclusion: It can be concluded that the relability and feasibility of exercise videos for LBP on YouTube depend highly on the upload source. Thus digital content in health should be created with a multi-disciplinary team including healthcare professionals and also standardized to inform patients with the risks and benefits to avoid the potential harm due to the complex causes of low back pain.