Inspiration

During the COVID-19 pandemic, online conferences and virtual meetings became essential for communication, education, and collaboration. However, language barriers made it difficult for people from different linguistic backgrounds to communicate effectively. This inspired us to develop Meet AI, a platform that uses AI technologies to make virtual communication more inclusive and accessible.

What it does

Meet AI is an AI-powered online conference platform that combines virtual meetings with multilingual communication and automated meeting assistance.

Key features include:

Real-Time Video Conferencing: Enables participants to communicate through virtual meetings.
Speech Recognition: Converts spoken words into text.
Multilingual Translation: Translates recognized speech into other languages.
Text-to-Speech: Converts translated text into spoken audio.
AI Meeting Anchoring: Provides automated speaker introductions, announcements, transitions, and time-related alerts.
Meeting Summarization: Generates a basic overview and extracts selected discussion points from stored transcripts.
Live Chat and Meeting Controls: Supports participant interaction and meeting management.

How we built it

We used React.js for the frontend and Node.js with Express.js for backend services. WebRTC supports real-time audio and video communication, while Socket.IO enables real-time communication between the frontend and backend.

For multilingual functionality, we integrated the browser’s Web Speech API for speech recognition, the Google Translate package for translation, and the Sarvam AI API for text-to-speech conversion. The anchoring service combines predefined announcement templates with translation and speech synthesis. MongoDB stores application data and meeting summaries.

Challenges we ran into

One of our main challenges was integrating multiple technologies into a real-time platform. Coordinating speech recognition, translation, and speech generation required careful handling of asynchronous operations and communication between frontend and backend services. Ensuring that announcements and translated content reached the appropriate participants was another important challenge.

Accomplishments that we're proud of

We are proud to have developed a platform that brings together virtual conferencing, multilingual communication, and automated meeting assistance. Integrating speech recognition, translation, and text-to-speech technologies allowed us to explore practical ways to reduce language barriers in online meetings. We are also proud of our teamwork and the opportunity to present our project at a hackathon.

What we learned

Through this project, we gained practical experience in full-stack development, real-time communication, API integration, and connecting AI services to a web application. We learned how WebRTC and Socket.IO support real-time communication and how speech-processing services can be integrated into a multilingual workflow.

What's next for Meet AI

We aim to improve translation accuracy and responsiveness, enhance meeting summaries, and develop more context-aware anchoring. We also plan to explore additional language support and improve the user experience to make virtual meetings more accessible to people from diverse linguistic backgrounds.