Thread receives a website URL as input from you to create a report for you based on that URL. After your submission, the URLs and their corresponding titles are placed in Thread's queue. When one of these URLs is picked off the queue, Thread uses an open-source library (Newspaper) to retrieve the contents of the website. Sometimes images are retrieved too but Thread focuses on the text it can retrieve from a URL (which is why you may have encountered errors when submitting some URLs).

With the text from a URL, this is then fed into machine learning (ML) models. These models have been trained with data to detect what tactics, techniques, and procedures (TTP) a sentence contains. Over time, we plan to improve the data the ML models learn from for better TTP predictions in sentences. As for now, after the ML models have done initial TTP predictions - mapping a sentence from the URL-text with a TTP - your report moves out of the queue and is ready for you to review.

If you are a website owner and you do not want us to fetch the contents of your site, you can email Arachne Digital via contact[at]arachne[dot]digital.