Background: Cognitive impairment is one of the most common and debilitating symptoms of multiple sclerosis (MS). Traditional neuropsychological assessment is time-consuming, only elucidates a snapshot of functioning at one time point, and takes place in an artificially controlled environment. Objectives: We sought to address these limitations by unobtrusively and continuously monitoring cognition in the patients natural environment with the innovative BiAffect application. Methods: BiAffect measures typing dynamics through a custom virtual keyboard. In this proof of concept investigation, 14 individuals with MS and 6 healthy controls (HCs) completed up to six weeks of data collection using BiAffect, which was installed on their personal iPhones. Participants also completed a battery of neuropsychological tests during a baseline assessment before starting to use BiAffect. The primary BiAffect metric is interkey delay (IKD), or latency between two consecutive keypresses on the BiAffect keyboard, which characterizes typing speed. Generalized multilevel models with Gamma distributions were conducted, with IKD during keypress sessions nested within individuals. Results: A total of 23,017 keypress sessions were collected. The MS group exhibited significantly longer IKD (or slower typing speed) than the HC group (estimate = -2.18, p < 0.001). Longer IKD was significantly correlated with worse performance on the Symbol Digit Modalities Test (SDMT; estimate = 0.06, p < 0.001) and other measures of processing speed, attention, and executive functioning. Memory performances were not significantly related to IKD. Conclusions: BiAffect is a promising, ecologically valid approach to remotely assess cognitive function, particularly abilities subserved by the frontal-subcortical brain system, among individuals with MS. Establishment of valid remote monitoring methods will enhance clinical management of cognitive dysfunction, which is imperative in a progressive condition such as MS.