Hey, you, get off of the cloud!
One of our goals with Talky is to make the service itself and its underlying components as secure as possible. For example, as Bear explained recently, we use a variety of industry-standard encryption techniques, including strong cipher suites to enable Perfect Forward Secrecy whenever possible.
Unfortunately, encrypting video conferences end-to-end with more than a few people is difficult. The challenge is that in peer-to-peer, “full-mesh” mode, your laptop or tablet or mobile phone needs to encode and encrypt one outbound video stream for each person involved. This works well for one-to-one video, but even the most modern computing devices simply can’t encode that much video data if you have more than a few people in the session.
To overcome that challenge, a service like Talky needs to use a server in the middle (e.g., the Jitsi Videobridge) that accepts one incoming video stream and fans it out to all the other participants. Unfortunately, this kind of Selective Forwarding Unit has to decrypt the video streams that it receives in order to do its job.