Background: Dietary intake is an important determinant of health. Its adjustments can improve the overall risk for noncommunicable diseases, especially for patients with obesity. Dietary assessment tools are commonly used to measure selfreported habitual energy intake (EI) and the overall diet. However, these tools are affected by voluntary and unvoluntary misreporting. Therefore, a preference for technology-based approaches has emerged. SNAQ is an image-based food recognition app. It has been developed to remotely measure dietary intake with a real-time transfer of data. Its validity in the measurement of EI has not been reported yet. The DLW methods is the gold standard for assessment of average daily energy expenditure (EE) in free-living conditions. It is commonly used to test the validity of other dietary assessment tools. Methods: Dietary intake was recorded for seven days with SNAQ in 30 study participants with and 30 without obesity. Urine samples for DLW analysis were obtained with a two-point protocol. Participants captured before and after pictures of each dietary item. Pictures were automatically uploaded, dietary items and portion sizes were estimated by a trained deep-learning model, and energy, macro-, and micronutrients were calculated. Paired t-tests, mean differences, linear correlations, and Bland-Altman limits of agreement were performed. Results: EI was mostly underestimated by SNAQ when compared to the DLW method. However, underestimation of EI by SNAQ correlated with the changes in body weight during the study week. Conclusions: We report a low validity of this method for the measurement of energy intake in free-living, healthy-weight adult women. SNAQ underestimated total EI when compared to the DLW estimations of EE. However, SNAQ seemed to perform better in participants with obesity. Further refinement of the use of SNAQ is needed for investigation of EI in free-living conditions.