A sound sensitivity layer for the internet — starting with a Chrome extension that makes movie night quieter.
There's no tool that lets you watch a movie without scaring your dog, waking your baby, or triggering your nervous system.
The technology exists. Google open-sourced a model that classifies 521 different audio events. Security companies use it. Smart doorbells use it. But nobody has pointed it at the thing most of us do every day: watch something on a screen.
A Chrome extension that sits between you and whatever you're streaming. It detects specific sounds in real time — doorbells, gunshots, screams, barking — and lowers the volume when it catches one, then brings it back up.
Toggle on the categories you care about. Adjust each one independently. Press play. That's it.
The audio never leaves your device. There's no account to create, no data sent anywhere, and no way for me to know what you're watching. The processing happens entirely in your browser.
I know what the exploitative version of this looks like — collect data on what bothers people, segment by condition, sell it. I designed the system so that's not possible. There's no user table, no viewing history, no demographic data. The architecture doesn't have a place to put that information even if someone wanted to.
The extension includes optional, anonymous feedback — a thumbs up if the filter worked, a thumbs down if it missed something. No account, no identity, just a signal.
Over time, those signals add up. The model gets better at catching the sounds that actually bother people. And the patterns that emerge — which sounds get flagged most, which moments in which films — could eventually become something useful for researchers, parents, and anyone who wants to know what they're about to watch before they press play.
That's the longer vision. But it starts with the extension just working.
Drop your email and I'll let you know when it's ready.