If you build an interview or reporting platform ask yourself where the raw footage goes. Who can access it. With GhostLine the answer is nowhere and nobody.
Hardest part of building GhostLine was not the anonymization. It was WebRTC. If you've debugged ICE candidates across browsers you know exactly what I mean.
No accounts. No logins. No stored data. Room exists for the call then it's gone. We don't want your data and we built it so we literally can't keep it.
Doing qualitative research where participants need anonymity? You can record the whole interview with GhostLine and never capture an identifiable frame.
Fun problem from last month: getting stable frame rates out of canvas captureStream while running the full anonymization pipeline without melting the CPU. OffscreenCanvas saved us.