Domestic violence is a public health issue that may be avoided for the most part and affects millions of individuals all over the globe. Without regard to factors such as age, race, or economic standing, it is believed that about one in every four women are victims of serious violence at some time in their life or have been victims of such abuse in the past. A significant number of victims choose to disclose their experiences with intimate partner violence (IPV) on social media platforms. The automated identification of such complaints through machine learning can enhance monitoring and facilitate the targeted distribution of assistance and interventions for those in need. However, there currently exist no artificial intelligence systems capable of automated detection of IPV. In this study, we aim to bridge this research gap by implementing a machine-learning model to detect cases of IPV on Twitter. Utilizing the Tweepy library and Twitter’s Streaming API, we engage in real-time data collection to identify tweets containing keywords associated with IPV as they are posted. We developed annotation criteria to categorize tweets as either IPV-report or non-IPV-report after manually reviewing subsets of these messages. Additionally, through the CTF-AIWF algorithm, we extract keywords related to domestic violence, ensuring a balanced data flow and allowing for a robust analysis of the discussions surrounding IPV on social media.