In modern digital communication, video conferencing has become an essential tool across various sectors; including work, education, and general communication. Traditional video conferencing systems have come a long way from their inception; however, they are yet to fully integrate artificial intelligence to provide necessary features such as real-time and post-meeting transcripts and automated summaries of the meeting. Hence, this study aims to modernize the video conferencing landscape by implementing features such as the previously stated real-time and post-meeting transcripts and automated emails containing the summary of the meeting. It still implements the core features required of any video conferencing application, such as the ability to create and schedule meetings, screen sharing, and audio and video communication. Natural Language Processing (NLP) models played a significant role in the implementation of the system through the Large Language Model (LLM) and Speech-To-Text (STT) models for the transcription of the meeting. In this study, Claude-3.5-Haiku and the OpenAI Whisper model were used for the LLM and STT models, respectively. The front end was done using React js and Tailwind CSS for a modern responsive user interface, while Python and FastAPI were used to manage backend functionalities, including database management and integration with the third-party APIs. The results highlight the effectiveness of integrating AI into video conferencing systems, improving user experience through enhanced automation and real-time processing. Future recommendations include further optimizing AI-driven features and expanding compatibility with additional third-party services to maximize system performance and adaptability.

Keywords: Video Conferencing, AI, Natural Language Processing, Large Language Models, Speech-to-Text, FastAP

Word count: 245