Reading proficiency is essential for academic success, where feedback is key for improvement. The level of feedback an educator provides has proven to be effective, however, the workload in the classroom hampers the timely feedback. NLP-powered tools have been demonstrating their effectiveness in providing adaptive and scalable feedback, similar to the feedback an educator provides. However, challenges remain related to adaptive and detailed feedback in the early stages of the development of reading skills. This research introduces the Reading Companion System (RCS), designed to fill these gaps by integrating adaptive feedback mechanisms and detailed pronunciation guidance. RCS dynamically scales vocabulary difficulty and offers personalized learning paths to maintain engagement and prevent cognitive overload. For pronunciation, it provides step-by-step phonetic breakdowns, enabling students to develop confidence and accuracy. We ran an evaluation study with 14 participants, including educators and caregivers of children aged 6–12, revealing that the system’s adaptive features are feasible and useful for vocabulary retention and pronunciation proficiency.