Hardware requirements

Our technology works without any extra hardware or apps. No VR glasses och other sensors are needed. Just join with your web browser or mobile phone and we transfer your emotional response by using the webcam.

Bandwidth requirements

Since no video is sent from or to your computer to our servers, the requirements for bandwidth is very low. An average video feed is between 1-2mbit/s which means that you would consume about 150mbit if you wanted to see 100 people at the same time. In our solution, each participant generates about 10 kbit/s of uncompressed data which means you can have thousands of users in your room without any problems and you can see them all in realtime.

Privacy

With great power comes great responsibility. We use AI and face expression recognition which should alert a sense of suspicion in you. We want to be very clear: we take privacy very seriously. To ensure your privacy we are taking these serious measures:

  • All data sent to our servers in Sweden are end-end encrypted using TLS and certificates issued by LetsEncrypt
  • No data is logged or saved on our servers. We have no data to sell.
  • All image processing and machine learning models is done client side.
  • No identifiable facial features or biometric data is sent to our servers.

Once we've found a sustainable business model (yes, we're a startup) we will publish our source code as open source. These are the technologies we use:

  • Face expression recognotion using Tensorflow mediapipe/facemesh.
  • Virtual Reality / Augumented reality using a-frame
  • Realtime communication using Node and Socket.io
  • Video Conferencing using Jitsi
  • Operations and scaling using Kubernetes
  • All of our servers are hosted in Sweden

Photo by Paul Hanaoka on Unsplash