Social media is used by everyone these days, although it is a great source for connecting people, it still has many drawbacks which cannot be overlooked. The big downside of social media is cybercrime. Great number of cybercrimes emanate from fake profiles. By detecting fake profiles, we can reduce cybercrimes. So, our project focuses on creating unified generalized Model to determine fake profiles over multiple social networks. We bring out this by training Artificial Neural Networks (ANNs), Random Forest model with fake profile datasets of different platforms such as Instagram, Twitter, Facebook. The model learns patterns and able to differentiate fraudulent profiles over real profiles. On continuous training and testing the model adapt multiple platforms and become generalized efficient tool that provide secure online space by combating phony accounts.