While purposeful disinformation, propaganda, and unscrupulous information are often spread on a select number of websites, these websites often rely on and are agenda-set by reporting from mainstream and otherwise reliable sources. Understanding the relationships between unreliable and reliable websites and how they cover the same topics in different ways is thus essential for understanding the spread of unreliable and false narratives. In this work, we introduce a system that by utilizing the daily scrapes of 18 months of news article data from 1,703 unreliable news websites and 2,374 authentic news websites, the large-language model E5, DP-Means clustering, and zero-shot stance detection, isolates and determines the relationships between unreliable and authentic news outlets and how these two types of websites cover different topics. We find that several mainstream websites such as Fox News are heavily influential in the unreliable news ecosystem and that unreliable news websites have tended to report anti-Ukraine, anti-vaccine, and anti-abortion information. We show that by comparing the stances of unreliable and authentic news websites and the volume of particular narratives, our system can both classify a website’s reliability and identify potentially misleading stories, thus helping the reporting and the fact-checking of misinformation.