Localizer — live tab captions by Lingo

Watch anything, understand everything. Localizer puts live captions over the video you're watching — generated on your own device, in real time, with optional translation built in.

How it works
Open a tab with a video playing, click the Localizer icon, and captions appear over the video — no built-in subtitles required. Set the source language, tick "Translate to English" for non-English content, and watch.

Features
• Live captions on any site — streaming services, news clips, lectures, webinars, anything with audio in the tab.
• Translation included — captions in the original language, or translated into English. No cloud translation; it happens on your device too.
• Rewindable captions — jump back in the video and the words are already there. Audio you've already watched is transcribed in the background into a rewindable caption timeline.
• Mute the tab, keep the captions — silence the video without stopping the captions.
• Remembered sites — Localizer can remember your language and translation settings per site and start automatically on your next visit.
• Fullscreen-friendly — captions render with your browser's native caption styling and stay on screen in fullscreen.

Privacy — yours stays yours
Everything runs on your device. The audio never leaves your machine. There's no account, no cloud service, no tracking, no analytics, no data collection of any kind. The only network traffic is a one-time download of the captioning models (around 200–3000 MB) the first time you use Localizer; after that, captions work offline. Nothing is transmitted to anyone.

Getting started
1. Open a tab with a video playing and click the Localizer icon.
2. Set the source language if the audio isn't English.
3. Tick "Translate to English" if you want translation.
4. Start captions.

Notes
• The first run downloads the captioning models, so captions begin a minute or two after your first click. Audio from before they load isn't lost — it appears in the rewindable timeline once ready.
• Captions appear a few seconds after the words are spoken (that's the price of on-device accuracy).
• One tab at a time.

Privacy form answers
• Data collected: None. No personal data, no audio, no usage data.
• Data transmitted: None. All processing is on-device; the only outbound traffic is downloading the captioning models, which is data received, not transmitted.
• Account required: No