In today’s digital world, cyber threats, especially social engineering attacks are becoming more sophisticated which is posing significant challenges to user security and data integrity. Social engineering includes various types of malicious tactics that exploit human interactions to bypass security protocols and gain unauthorized access to sensitive information. To tackle these threats, we need advanced detection systems that can identify subtle signs of deceit in real time. This paper presents a new detection system that uses BERT, a leading natural language processing model, to combat social engineering threats. Our system taps into BERT’s strong capabilities to recognize deceptive language patterns commonly found in social engineering attempts. By integrating this system into a Flutter mobile application, it quickly alerts users to potential threats, enabling them to make informed decisions and protect their digital assets. By detecting deceptive tactics swiftly and accurately, our proposed system enhances users’ ability to resist evolving cyber threats, addressing the urgent need for innovative security solutions in today’s digital landscape.