For information security, certain vulnerabilities can lead to major effects which later can cause consequences for economic, social, and national security. It's predicted that through the use of top-notch NLP technologies in the field of vulnerability detection, a model can be created for the accomplishment of the automatic analysis which will help in the detection of particular text files such as source code. An in-depth learning-based vulnerability detector is proposed that pulls out some features through the help of an RNN composite neutral network. They employ a SARD and NVD dataset of lots of open source roles, which will be tagged with results after three static analyzers that hint at possible activities. Through the implementation of this data set, a fast and accessible detection system will come into being with built-in deep feature sign knowledge that will directly translate the source code. Based on our research, we found out that through deep feature representation learning on source code the software vulnerabilities can be automatically unrevealed. In this study, the model vulnerability detection is used as a natural language processing (NLP) problem where source codes are handled as texts. Furthermore, advanced deep learning NLP models are there which offer transfer learning on written English to address/look into automated software vulnerability detection. In short, we have used different algorithms i-e, machine learning, and deep learning.