I've set up more interpreter booths than I can count.

I've spent years as a sound engineer and AV technician, mostly on events that needed simultaneous interpretation. The same pattern played out at almost every one of them: an interpreter booth, IR radiators, a console, a box of receivers, plus delivery, installation, and teardown - even for an audience of ten people or fewer.

The equipment is the expensive part, and it doesn't scale down. A booth costs roughly the same whether you're translating for two hundred people or twelve.

Early AI translation was rough to listen to live - the robotic delivery made people tune out fast. That changed with newer models trained specifically for simultaneous interpretation. It stopped being something I'd be embarrassed to offer a client.

Human interpreters are still better in plenty of situations - they can catch a presenter's mistake, read between the lines, or skip a tangent that doesn't need translating. Where AI wins is fast, dense speech, which is exactly where human interpreters struggle most.

Getting translated audio to a listener's phone over WiFi with no perceptible lag, and with no app to install, was the real engineering problem. Boothless solves it with a QR code and a browser tab.

I still provide simultaneous interpretation for my own clients, and Boothless is the tool I built to do that job better and more cheaply - for myself, before anyone else.

Every user brings their own Gemini API key, so translation is billed at Google's real cost, with nothing added on top.