For the languages that are not widely spoken, such as Kazakh, it is highly important to have digitized and processed handwritten text in order to maintain linguistic diversity and have rich digital resources. The purpose of creating this article will be the development of a Kazakh advanced-handwritten text recognition system using the most effective cutting-edge methods of deep learning. The current system mainly provides the capability to identify and convert digital use of handwritten Kazakh text. It started with the collection of a large diversified dataset. The data was collected from students of Astana IT University and is to represent wide diversity in handwriting styles. Procedures were then carried out to scan the high-resolution handwritten images and trim each character manually from the scanned images. The dataset was divided into a training and validation set to enable effective training and evaluation. Some preprocessing techniques such as grayscale conversion and normalization were applied, which were necessary to improve the quality of the images before training. The strategy of the proposed recognition is a hybrid deep learning approach, achieved through Convolutional Neural Networks (CNN) to extract features and Recurrent Neural Networks (RNN) by way of Long Short-Term Memory (LSTM) for sequence modeling.