Document accessibility refers to designing and implementing documents that can be easily accessed and understood by persons with disabilities, ensuring equal access to information and services. Document summarization plays a vital role in improving document accessibility by providing brief but informative summaries that allow users to quickly understand the primary information of a document, particularly for individuals with limited time, attention, or cognitive abilities. It involves extracting key terminologies or generating a summary that provides an overview of the document’s essential concepts. This study proposes a novel system for improving document accessibility for people with visual impairment that uses advanced natural language processing techniques. The design includes two key components: keyword extraction using the TF-IDF algorithm and document summarization using the GPT-3 language model. By combining these two techniques, the system enables users to quickly identify relevant documents based on extracted keywords and obtain brief summaries with text-to-speech features that enable effective understanding. The proposed system achieved a remarkable average F1 score of 0.53, 0.36, and 0.51 for ROUGE-1, ROUGE-2, and ROUGE-L, respectively, for summarizing the scanned documents.