Integrating AI and Encryption in WebRTC: A Smart Video Conferencing Platform with Real-Time Transcription and Translation

Integrating AI and Encryption in WebRTC: A Smart Video Conferencing Platform with Real-Time Transcription and Translation

Samarth Patil1, Vinayak Nigam2, Shriyash Olambe3  and Ms. Trupti G.Ghongade4

1Samarth Patil, Computer Engineering, Marathwada Mitra Mandal's College of Engineering

2Vinayak Nigam,Computer Engineering, Marathwada Mitra Mandal's College of Engineering

3Shriyash Olambe, Computer Engineering, Marathwada Mitra Mandal's College of Engineering

4Ms. Trupti G.Ghongade,Computer Engineering, Marathwada Mitra Mandal's College of Engineering

---------------------------------------------------------------------***---------------------------------------------------------------------

Abstract - As the demand for seamless, real-time communication tools grows, traditional video conferencing platforms need help to balance simplicity with advanced features. BabelTalk aims to address this challenge by integrating AI-powered capabilities into a scalable, secure, real-time communication platform built on WebRTC. WebRTC relies on basic peer-to-peer architectures, which limit scalability and also lack pre-processing of media streams, BabelTalk takes inspiration from a quasi-peer architecture to enable advanced features such as real-time transcription, translation into captions or speech, and post-meeting summaries while ensuring secure media transmission. The platform prioritizes data privacy and offers a feature-rich communication experience tailored for small to medium-sized teams by utilizing isolated Dockerized services for temporary decryption during AI processing.

Key Words:  WebRTC, synchronization mechanism, media synchronization, congestion control, Speech-to-speech translation (S2ST), Text-to-speech synthesis (T2S), Unit-to-unit translation (UTUT)