FEATUREDLatestOpenSIPSTOP STORIESVOIPWebRTC

WebRTC Security: How Safe Are Browser Calls?

WebRTC lets people make voice and video calls straight from a web browser. As more businesses use it for customer and team calls, the obvious question is whether it is secure. The short answer: the standard itself is strong, and most risks come from how it is deployed.

Encryption is built in, and mandatory

Browsers will not send unencrypted audio or video over WebRTC. The WebRTC security architecture requires it. Three parts do the work:

  • SRTP (Secure Real-time Transport Protocol) encrypts the audio and video.
  • DTLS (Datagram Transport Layer Security) is used to agree the encryption keys for SRTP, a method known as DTLS-SRTP, and to encrypt WebRTC data channels.
  • ICE, with STUN and TURN servers, finds a working network path between the two ends. ICE also checks that the other side has agreed to receive media (consent freshness), so a web page cannot use your browser to flood someone else with traffic.

For a deeper comparison, see DTLS-SRTP vs SRTP for PBX admins and Siperb’s explainer on DTLS.

Signalling is your responsibility

WebRTC does not define how calls are set up. That signalling is left to the application, which usually carries it over HTTPS or secure WebSockets (WSS), often as SIP over WebSockets. It matters because the fingerprints that identify each side’s encryption keys travel in the signalling. If an attacker can tamper with the signalling, they can try to place themselves in the middle of the call. Always run signalling over TLS.

What about IP address leaks?

Early WebRTC implementations let any web page discover a visitor’s local network address. Modern browsers, including Chrome and Firefox, now hide local addresses behind random mDNS names unless the user has given the page camera or microphone access, following the rules in RFC 8828.

Your public IP address is still visible to the servers you connect to, as it is for any website, and to the other party on a direct peer-to-peer call. If that matters, an application can force all media through a TURN relay so the other party only ever sees the relay’s address. Disabling WebRTC in the browser also works, but it stops browser calling altogether.

Where encryption can stop: gateways

When a browser call connects to an ordinary SIP phone system, a gateway often converts the browser’s DTLS-SRTP into the plain RTP or SRTP the PBX expects. At that point the audio is decrypted, so the call is no longer encrypted end to end. Siperb explains how its calls stay encrypted end to end when no conversion is needed. Applications that need end-to-end encryption even through media servers can add it with insertable streams.

Is WebRTC safe for business?

Yes, when it is implemented well. The weak points are usually around it rather than in it: unpatched softphones and browsers, weak SIP passwords, signalling that is not encrypted, and open international dialling that invites toll fraud. For a hands-on example, Innovate Asterisk shows secure WebRTC calling with an Asterisk PBX.

Related reading

Updated 11 November 2026: corrected what DTLS does, updated the section on IP address leaks for current browsers, and added gateways and signalling.

Leave a Reply

Your email address will not be published. Required fields are marked *