WebRTC is the technology behind in-browser video calls, voice chat, and screen sharing that don't require installing a plugin or separate app — a genuinely useful piece of modern web infrastructure, and one with some privacy nuance worth understanding.
What WebRTC actually does
WebRTC provides browsers with built-in APIs for capturing audio/video, and for establishing a direct, low-latency connection between two peers (or via a relay when direct connection isn't possible) without needing dedicated server infrastructure to carry the actual media stream.
Signaling: setting up the call
Before two browsers can connect directly, they need to exchange some initial information — this is called signaling, and it's typically handled by whatever application is using WebRTC (a video call service, for example), through its own server, using whatever protocol it chooses. WebRTC itself doesn't standardize signaling; it's the connection-establishment phase, not the actual media transport.
ICE: finding a path between two browsers
Once signaling has exchanged basic information, WebRTC uses a process called ICE (Interactive Connectivity Establishment) to figure out the best actual network path between the two peers — trying a direct local connection first, then various fallback options via STUN and TURN servers if a direct path isn't available (common when both sides are behind NAT, which is most home and mobile connections).
Why no single server carries every call
WebRTC's design goal is a direct peer-to-peer connection wherever possible, rather than routing all media through a central server — better latency and lower server costs for the application. This same address-discovery process, though, is also the mechanism behind WebRTC leaks — see WebRTC Leaks Explained.
FAQ
Do I need to install anything to use WebRTC?
No — it's built into modern browsers by default, which is the whole point of the technology.