Loom

Dual subtitles with romanization for foreign-language video on YouTube, Netflix, Prime Video, iQIYI, and WeTV.

Loom: Dual-language subtitles for YouTube, Netflix, Prime Video, iQIYI, and WeTV with phonetic annotations, romanization, and a per-word pop-up dictionary. Loom is a language-learning tool for anyone who watches YouTube, Netflix, Prime Video, iQIYI, or WeTV in a language they're studying. It renders TWO subtitle tracks at once (the original (foreign) language on top, your own language on the bottom) so you can follow the dialogue without switching back and forth. For non-Latin scripts, Loom adds a romanization line above the foreign text (eg "konnichi wa" above "こんにちは") so you can read along even before you learn the characters. For CJK languages it additionally adds per-character readings: furigana for Japanese, Pinyin or Zhuyin for Chinese, Jyutping for Cantonese, Revised Romanization for Korean. For Traditional Chinese, Loom can also show the Simplified form of each character above it. NEW: Per-word pop-up dictionary, now across ~20 languages. Pause the video and click any word in the top line: a card shows its reading (furigana for Japanese, Pinyin for Chinese, Revised Romanization for Korean), its dictionary definition, and - for Japanese and Korean - a grammar breakdown that traces an inflected word back to its dictionary form (食べさせられた → 食べる · causative · passive · past; 가셨어요 → 가다 · honorific · past · polite). Japanese words also get Hepburn romaji (e.g. "Tōkyō"). Loom breaks each sentence into real words — grouping multi-character words, recovering a verb's dictionary form, and handling measure words and name honorifics — so your click lands on the whole word, not a stray character. You can even switch the language the definitions are written in. Definitions come from the JMdict, CC-CEDICT, KRDict, and Wiktionary open dictionaries. Loom also keeps each caption where the show puts it: vertical text, on-screen signs, and stacked simultaneous lines render in their original position and orientation instead of being flattened into a single row. Every visual aspect is customizable: per-layer colors, fonts, sizes, outlines, glow, and opacity, plus a per-site subtitle-size and vertical-position adjustment. 28 thematic color presets ship out of the box. Settings persist per device. Loom works on any video that has a selectable text subtitle track in your target language (on YouTube, that means human-made or auto-generated captions). Loom is a research project from nerv-analytic.ai. It's free, has no ads, and collects no personal data. The only thing it sends out of your browser is subtitle text (used to generate the romanization, and (when you click a word) to look up its definition). You can also choose to contribute subtitle text to help improve Loom's quality; this is off by default and asked for once. See the privacy policy at loom.nerv-analytic.ai/privacy for details.