Modern computers can not only record us but analyze what we say. Online privacy is a growing concern as more and more smart devices record our speech, and there is no reliable way protect our personal information from potential eavesdroppers.
What it does
We utilized properties of sound to ensure that only the intended recipient is reached. The sender's speech is distorted, and Pulsar generates and sends the recipient a noise that cancels out this distortion.
This ensures that only recipients with the right information can make any sense of the sender's speech. Our RESTful API allows developers to easily implement this system on any web-based application and protect their customers' private information. Pulsar means that people can have private conversations with each other without fear, and more importantly, it ensures that no third parties can steal private information as audio travels over the web.
All of the audio processing happens on the client-side, so not even Pulsar has access to users' audio.
How we built it
We used node.js and python for the back-end, and we deployed it with Heroku. On the front-end, we used Materialize and Sass. We used audio processing libraries such as tone.js and record.js as well the web audio API for everything audio-related.
Challenges we ran into
-Getting the many sounds to play exactly when they should without latency -Processing audio to achieve noise cancellation -Coming up with a suitable way to distort audio