Telegram is used daily by millions of people to communicate. It contains a multitude of discussion topics, some of which are violent, illegal, or extremist in nature. To monitor public chats, various crawlers have been developed. Their main problem is their ineffectiveness in finding channels of interest in a timely manner based on a search topic, as well as their inability to retrieve all groups of interest. The following thesis aims to overcome these limitations by understanding whether it is possible to develop a crawler capable of not only obtaining almost all groups of interest, but also doing so using limited resources and in a reasonable amount of time.