There are a lot of video chat applications in existence. However, most are plagued by complex interfaces, messy authentication systems and platform dependencies that hinder a groups ability to get together. We wanted to build something to fill the void.
What it does
Meet is an intelligent WebRTC based video conferencing system which requires no additional software or even authentication. Provide an email, name or both and in under a second, you'll be running your own chat room, private to those you choose to invite. Invite your colleagues and friends via simple URL invitations. Meet also employs the use of speech recognition technology, and jots down light notes as conversations occur.
How we built it
We built Meet using WebRTC, Node, React, Express, Redux and the Google speech API.
Challenges we ran into
We ran into a number of issues regarding speech analysis, particularly the streaming of system audio to an analyzation service.
Accomplishments that we're proud of
We're quite happy that we were able to produce a quality, stable application within the time confines. Overall, the application is very stable and works reliably.
What we learned
We both learned more intricate details of WebRTC than we ever thought possible. From TURN to STUN to ICE and everything in between, we dabbled through the WebRTC stack and ever ended up learning a bit about audio processing.
What's next for meet
Meet's intelligent capabilities can be enhanced and refined far beyond what we currently support. Additionally, we'd like enhance our email reporting to include additional call metrics.