Parkinson’s disease frequently manifests with vocal symptoms, such as monotonicity, breathiness, and irregular pitch, all correlating with motor impairments and detectable through speech features. Advances in signal processing and machine learning have made these vocal markers a promising tool for noninvasive, cost-effective screening. However, several challenges limit the clinical adoption of these tools. Limited data and the higher focus on English-speaking populations, raise concerns about model’s performance across languages. In low resourced languages, this can hinder the clinical validation of such opportunities. This paper examines the development of resources and tools for PD voice-based characterization in European Portuguese over the last two decades. Beyond diagnostics, it explores the potential of vocal biomarkers as real-time feedback mechanisms for PD management. Finally, it addresses the ongoing challenges and limitations to achieving universal adoption of voice-based PD diagnostic systems.