Small changes.Better learning.

Understand the word. Connect the idea.

Word explanations show the base form and its use in the sentence. Knowledge maps explain key ideas and clearly distinguish suggested connections from the video’s own content.

• Older cached word highlights can receive base forms while you read, without replacing your highlighted passages. Uncertain forms are not requested repeatedly.
• Knowledge keywords open the exact source passage. Existing maps remain readable, with an explicit option to update them.
• Sentence preparation handles Japanese and Chinese captions more carefully, limits repeated AI calls, and keeps long fallback sentences readable.

Learn from what is on the screen

Capture a video moment for a screenshot lesson. The website also shares learning data with the extension through your account.

• Screenshot lessons keep the image, video moment, explanations and practice together. Capture and preparation errors offer a way to retry.

Open a video and start reading

All ten homepage lessons include prepared bilingual sentences, highlights, outlines and knowledge maps. Reading is open without an account.

• Paste a featured video link to open its free lesson directly. Other videos continue into your private learning space after sign-in.
• Bookmarks stay attached to the correct lesson edition, and knowledge-map selections survive reading-progress updates.

Follow an idea into the next video

Explore a video’s knowledge points, trace their source evidence, and look for videos that answer your next question.

A first lesson, without the friction

Open a prepared lesson without signing in. Read, save a thought, and try a short recall round at your own pace.

• Ten homepage videos open prepared full English transcripts without an account. Sign up when you want to bookmark a word, passage, or note. The solar-panel lesson also offers five explanation languages.
• Choose your explanation language separately from the interface. Unsupported language requests stop early and offer a way to submit the video URL.
• Visible word interactions, clear note-saving feedback, preparation progress, and links between saved notes make the learning loop easier to follow.

Your videos, connected to your practice

Knowledge becomes a video library, and practice gets a focused space of its own.

• Resume watching, read the full transcript, organize video tags, and practice from selected videos—even before saving an excerpt.
• An excerpt companion answers within the nearby source context. Public email signup was also repaired so new learners can enter the workspace.

Explanations that stay closer to the source

Vocabulary and grammar explanations receive stronger source checks and better recovery when generation is interrupted.

• Repeated explanations and duplicate grammar forms are filtered. Selecting a whole sentence now asks for useful vocabulary within it.

Start with something worth watching

Real videos become the entry point, with 50 checked recommendations and a homepage built around a playable excerpt.

• Sentence reconstruction, original-audio dictation, and rewriting keep practice tied to the video you chose.

More ways to find your way in

Public learning guides introduce the video-learning workflow in more languages.

A reading room for the things you watch

A paper-like reading workspace brings complete sentences, bilingual explanations, and learning highlights together.

• Long-video preparation is repaired, repeated preparation work is reduced, and saved excerpts retain their source context.

Video Sensei finds its own home

The independent website brings sign-in, video learning, saved knowledge, and practice into one connected flow.

• Private workspaces store learning records, while bilingual translations stay anchored to exact source fragments.

Watch, understand, save, return

The earlier prototype connects its learning loop and refines the reading layout and phrase alignment.

A workspace that remembers the video

The prototype is rebuilt around video-based learning journeys, a persistent workspace, and query-driven YouTube search.

Beyond reading: audio and quizzes

Early experiments add audio, tailored summaries, multilingual support, and quizzes to the transcript experience.

Making transcripts easier to work with

Transcript fetching and analysis are separated, and smaller chunks improve processing performance in the first prototype.

It started with a video

The project begins. In the following days, the first YouTube integrations and transcript experiments take shape.