Source-attached Korean comprehension

Interlign

Interlign keeps the Korean line visible, explains its words and grammar in context, and turns the exact media moment into review material.

Watch demosGloss textBuild a video pack

The demos and text glosser are open. Video packs are the downloadable path for longer material.

VideoGloss real clips, with transcription when usable subtitles are missing.
SenseResolve the sentence meaning instead of dumping dictionary guesses.
MiningSave the exact moment with screenshot, audio, translation, and gloss.
Korean video frame with the Interlign cinematic split gloss overlay, C-view grammar insight, roots, display markers, and plain translation.
Interlign overlayReal product capture: source line, split gloss, C-view grammar, roots, particles, and plain translation stay on the same media moment.
Korean liveMorpheme viewNo captions pathReview card mining
Watch the product work

Native video, glossed in place.

These clips are generated by the current cinematic split renderer, not decorative mockups. Source line, gloss layer, C-view grammar rail, roots, display markers, and plain translation are all enabled in the export.

Clip from SemoDuck's Rusty Lake video "The Mr. Rabbit Magic Show." Watch the full video on YouTube.

Cinematic split
C view + plain translation

The current split renderer keeps Korean, gloss, grammar insight, roots, particles, and plain translation visible.

Cinematic split
Particles + modifier structure

A second real Korean clip shows display markers, root labels, and the C-layer on another line.

Evidence reel
From real media to a saved card

Press V on the line you are watching. Interlign keeps the source frame, audio, Korean, gloss, and translation together.

Tap V on the lineThe card starts at the exact media moment.
Keep the source frameScreenshot and audio stay attached.
Auto-build the backGloss, translation, and context travel with it.
Send to SRS or AnkiReview starts from real input, not isolated vocab.
Interlign cinematic split gloss overlay, C-view grammar insight, roots, display markers, and plain translation on a Korean video frame
Context glossThe Korean line stays visible while the gloss, grammar note, roots, particles, and plain translation stay attached to that moment.
Interlign cinematic split gloss overlay on a Korean video frame before mining
Source momentMining starts from the actual frame being watched, not a detached vocabulary row.
Interlign review card made from the same Korean video moment with the source screenshot attached
Mined cardThe saved card keeps the frame, Korean line, gloss, translation, and audio together.

Better than a pop-up dictionary: Interlign uses the sentence context to expose endings, auxiliaries, constructions, and discourse stance that candidate-word lookup often flattens.

The sentence decides the gloss.

Popup lookup is good at listing candidates. Interlign is built to keep endings, auxiliaries, stance, and constructions attached to the words that produced them.

Interlign reads 떠나야지만 알 수 있는 게 있죠 as leave-must-only know can existing thing exist-right while lookup misreads 알, 수, and 게.
Interlign reads 그렇게 찢어발기지 좀 말라고 좀 as like-that tear-rip-open-don't please don't-said please while lookup loses the negated reported command.
Interlign reads 감싸고 돌잖아 as shield-and go-around-you-know while lookup makes it look like physical wrapping and rotating.
Surfaces

Video is only the first lane.

Interlign is not limited to video. The same glossing model is being pushed across books, games, apps, browser pages, and anything else with target-language text.

Game OCR

Prototype capture: game OCR is still under active development.

Reader

Extension preview: the same source-attached gloss pattern transfers to reading and longer text.

Browser extension

Extension preview: browser reading keeps the source page intact while adding compact word-level support.

Public comments