The article discusses automated keyword definition in text messaging for social networks that analyzes text message using the methods of TFIDF estimation, variance estimation, and TFIDF estimation using NLP. The developed automated keyword definition information technology was implemented in the test software. The input data of the system is a text message with digital text, and the output data is a text message with a set of key terms. In developing a socially-oriented interest-based communication service on the iOS platform, according to the defined functions, the following groups of users are identified: registered user, social network administrator, unregistered user. A registered user works with the system via the IOS Mobile App – this group includes users who are logged in and have access to photo and video messaging, location sharing, search for other users, viewing other users’ news feeds, commenting on news from others users, create their own news feed, view their own news feed, and track other users’ news. The social network administrator works with the system through a browser interface on an arbitrary platform – this group includes users whose function is to backup the database, view the news feed of any user, edit all spreadsheets, exchange messages, lock the user and password reset. The unregistered user only has the option to register. With the help of the developed test software, studies were carried out, which confirmed the possibility of effectively automated determination of a set of keywords in text messages with accuracy indicators for the TFIDF estimation method – 27,1% and the dispersive estimation method – 45,5%, the TFIDF estimation method with NLP 88,3%. Advantages of the developed information technology of automated definition of keywords in text messages for social networks, which conducts the analysis of text message using the methods: TFIDF estimation, variance estimation and TFIDF estimation using NLP, there is no need to use lexical databases of corpora of words, significant acceleration possibility to use for texts in different languages, possibility to use for texts with several languages. This information technology can be effectively used to analyze text messages with unknown subject and language properties.