Every day, big amounts of unstructured data is generated. This data is of the form of essays, research papers, speeches, patents, scholastic articles, book chapters etc. In today's world, it is very important to extract key patterns from huge text passages or verbal speeches. This paper proposes a novel method for summarizing multilingual vocal as well as written paragraphs and speeches, using semantic Knowledge Graphs. Using the proposed model, big text extracts or speeches can be summarized for better understanding and analysis. The method uses speech recognition as well as Named Entity Recognition to identify entities from spoken content to create optimized Knowledge Graphs in the English Language. © 2019 IEEE.