This practice led me to wish for a feature that would automatically pause a video at the end of each subtitle. This pause would provide me with time to:

- Process what I just saw and heard.

- Mine words using Yomitan, if necessary.

- Replay the segment to shadow the speaker, ideally triggered through a keyboard shortcut.

- Analyze the sentence structure for deeper insight on complex grammar and cultural nuances.

- Translate the entire sentence in a language I know, if necessary.

To enhance my language learning experience I developed a Chrome Extension called "Subtitle Insights". This extension leverages Chrome's Built-In AI (specifically Gemini Nano) to perform on-device translations and analysis of YouTube subtitles.

Key features:

- On-Device AI Processing. Once Gemini Nano is downloaded by Chrome, all subtitle translations and language insights are processed locally using Chrome's integrated AI capabilities.

- The prompt is customizable and the output can be tailored to match your preferences. If you're learning multiple languages at a time you can create profiles for each one.

- The auto-pause feature pauses the video just before a subtitle segment ends and gives you time to fully process the spoken content.

- The sidebar displays all subtitles and can be used as a video navigation tool. You can jump to any subtitle with a simple click.

- Keyboard shortcuts allow for quick navigation between subtitle segments and easy replay of the current segment. You can bring your own subtitles and if they're not synced with the audio you can use the extension to sync the audio with the subtitles easily.

It's free. No account / API keys needed. The Gemini Nano model runs on-device thanks to Chrome's Built-in AI.