This paper presents an end-to-end application leveraging transformer-based architectures for Vietnamese handwritten text recognition (HTR). Handwritten text presents unique challenges due to variability in writing styles, noise, and distortions, especially in low-resource languages like Vietnamese. The system is designed to preprocess input images using Contrast Limited Adaptive Histogram Equalization (CLAHE) and Sauvola thresholding to enhance text visibility. Subsequently, the preprocessed images are passed to a transformer-based model, which captures both spatial and contextual information from the handwritten text. To evaluate performance, we benchmarked the system using a custom dataset of Vietnamese handwritten documents. The results demonstrate that the transformer-based OCR system achieved great accuracy and robustness across various handwriting styles test in our custom data with complex background. This application aims to contribute to the growing field of HTR in Vietnamese, providing a scalable solution for digitizing historical documents, forms, and other handwritten content with promising results of 9% and 24% for Word Error Rate and Sequence Error Rate, respectively.