SocketTranscriber

Inspiration

I wanted to use the Assembly AI transcription API and learn more about web sockets

What it does

It is a real-time socket based transcription service. The users are split up into rooms where each room has a speaker who as they speak gets their audio transcribed into text and then sent to the listeners of the room in real time.

How we built it

I first brainstormed some ideas using the API to see how it works. I then created a few quick diagrams to make sure the architecture was solid. After that I used the Assembly AI API, a basic Express web server, web sockets, EJS, WebRTC, and Docker to create the app.

Challenges we ran into

I tend to focus more on backend applications so there was a bit of Googling that had to be done to get the frontend modules working.

Accomplishments that we're proud of

It was cool to finish my initial commitment to the project and have it look decently nice (at least for my standards).

What we learned

I learned about WebRTC and its always fun to learn a new cool API.

What's next for SocketTranscriber

I started to deploy it on AWS but HTTPS is needed for WebRTC and I didn't have time to add a load balancer that could handle the secure connection, so I'd like to add that so I could fully deploy the app instead of running it locally.

Built With

• docker
• ejs
• express.js
• node.js
• webrtc
• websockets