Despite the prognosis that the aging population is expected to have on increasing stroke prevalence, current advances in the availability of digital technologies for aphasia rehabilitation provide hope for those affected by this condition. The VoiceAdapt project aims to empower elderly people with aphasia (PwA) by means of improving their communication capabilities and their daily communicative activity to attain greater levels of long-term recovery. To this end, the consortium applies innovative speech-sensitive technology to application-based support & training of PwA in order to develop a system that detects and adapts to spoken language deficiencies as typical signs and symptoms of aphasia. Building on existing speech training and adaptive testing approaches, at the core of the system is an intelligent personalization and adaptation engine, supported by gamification mechanisms, that receives input from multiple data sources (language training application, voice and interaction pattern analysis) and dynamically adapts a) the multimodal interface of the application as well as b) the behaviour itself to individual mental states and cognitive capabilities, in particular with regard to their cognitive disabilities. Continuous involvement of PwA, their caregivers and medical experts during development leads the way to the conduction of a randomized controlled trial (RCT) with PwA to measure and monitor the impact of the developed adaptive training tool on participants´ linguistic abilities, communication habits and quality of life.