Free tools to make a WebRTC dump safe to share
You will find here a few open source tools that you can use to process your down webrtc-internals and rtcstats dump files. These can be useful before analyzing files or sharing them publicly.
WebRTC dump anonymizer
“I need to attach this dump to a bug report, but it has our users' IP addresses in it.”
Drop a webrtc-internals or rtcstats dump on the page and the anonymized file downloads. It rewrites host and server-reflexive ICE candidates, the a=candidate: lines inside the SDP, and the IP addresses inside getStats() records. TURN server and relay addresses stay as they are, so the call flow still makes sense to whoever reads it.
For example, 192.168.1.42 becomes 192.168.1.x. For IPv6 only the network prefix is kept. You can still see that two candidates sat on the same subnet, or that traffic went through a specific TURN server.
It covers network identifiers only. Values you set yourself, like userId, sessionId or device names, stay in the file. See exactly what it changes.
WebRTC dump splitter
“The dump has every tab I had open. The vendor only needs the call from our app.”
webrtc-internals records every page in the browser that used WebRTC. The splitter reads the dump, lists each origin with its peer connections and getUserMedia calls, and exports only the origins you tick. Timestamps are kept as they were.
For example, you tested your app at app.example.com while a Google Meet call ran in another tab. The dump shows two origins. Keep yours, download, and the Meet connections are gone from the file.
It accepts webrtc-internals JSON dumps and rtcstats dumps (.json, .jsonl, .txt, .gz).
Using both? Split first
Cut the dump down to your origin, then run the result through the anonymizer. What is left is your call, without anyone else's IP addresses or tabs.
Want the answer instead of raw JSON? Upload the dump to rtcStats and get Observations, an Experience Score and an AI Summary. The free plan covers it.
Both tools are open source, in the rtcstats repository on GitHub.