Parkinson’s disease is a chronic neurodegenerative disorder that progressively affects motor function. One of its most characteristic symptoms, bradykinesia, is typically evaluated through clinical observation, which introduces subjectivity and inter-observer variability. Current measurement methods are subjective, time-consuming, and reliant on clinician expertise, limiting consistency and accessibility. This paper presents the design, development, and implementation of a computer vision and artificial intelligence–driven system, delivered via a web-based platform, to automate and digitize the clinical evaluation and follow-up of Parkinson’s patients. Grounded in the criteria of the MDS-UPDRS Part III, the system enables objective and real-time measurement of the finger-tapping task. The platform extracts parameters such as tap frequency, opening and closing velocity, amplitude, and hesitations, using video analysis powered by AI. These metrics are stored in a structured database, allowing for visualizations, patient progress tracking, and remote clinical monitoring. Additionally, the system supports comparative evaluations of patients under different medication states, reduces inter-observer bias, and provides consistent criteria to assess disease progression. The result is a reproducible, low-cost, and scalable digital tool designed to assist neurologists and researchers in delivering more accurate diagnoses and continuous patient follow-up.