Primary brain tumours frequently lead to speech impairments, with disease progression rates remaining high despite intensive multidisciplinary treatment. Existing diagnostic tools often lack the sensitivity and specificity required for effective monitoring, highlighting the need for additional biomarkers. Speech has emerged as a promising candidate, as its disruption may reflect disease progression in neuropsychiatric conditions such as amyotrophic lateral sclerosis, multiple sclerosis and Alzheimer's disease. However, research in this area remains limited, and no dedicated speech corpora currently exist for this population. To address this gap, we launched The BrainApp Study, designed to remotely collect longitudinal speech, clinical, physical activity, and quality-of-life data from individuals with primary brain tumours. To date, we have curated 285 intelligible read speech audio samples—186 from 27 patients—with accompanying manual transcriptions. Group-level differences were observed in 15 of 20 acoustic features between patients and healthy adults. These preliminary findings highlight both the feasibility of large-scale, remote speech data collection in a population with high morbidity and limited long-term survivability, and the potential to capture clinically salient speech features in patients. This can lay the groundwork for future digital speech studies in neuro-oncology.