Background: Alzheimer's disease (AD) dementia and Mild Cognitive Impairment (MCI) are currently underdiagnosed in the community. Early detection of cognitive deficits is crucial for timely intervention. FACEmemory®, the first completely self-administered online memory test with voice recognition, has been launched as an accessible tool to detect such deficits. This study aims to correlate FACEmemory subscores with traditional neuropsychological tests and to develop an algorithm to distinguish between cognitively healthy (CH) individuals and those with MCI. Methods: A total of 669 participants (266 CH, 206 with non-amnestic MCI -naMCI-, and 197 with amnestic MCI -aMCI-) were included. Pearson’s correlations were conducted to examine the association between FACEmemory scores, demographics, and cognitive performance on classical neuropsychological tests (the Neuropsychological battery of Ace -NBACE-). Machine learning models were compared to find the best algorithm for distinguishing between CH and MCI (whole MCI, aMCI, and naMCI). Results: FACEmemory scores showed significant correlations (p< 0.001) with memory (r= 0.53), executive functions (r= 0.43), visuospatial/visuoperceptual (r= 0.32), language (r= 0.30), praxis (r= 0.30) and attention (r= 0.26) domains of the NBACE. The best algorithm distinguished between CH and MCI (especially aMCI), with a FACEmemory cutoff score of 44.5, yielding sensitivity and specificity values of 0.81 and 0.72, respectively. Conclusions: FACEmemory is a promising online tool for detecting MCI, particularly aMCI. It could play a key role in reducing underdiagnoses of MCI and AD and supporting early intervention efforts in the community.