KitsumiKitsumi itself is proprietary; see LICENSE in the project repository at https://github.com/Olexify/Kitsumi. This file covers the parts of this package that are NOT Kitsumi's, and it exists because their licences require their notices to travel with the code. Those components keep their own licences: the terms below govern them, not Kitsumi's terms.
Every full licence text named here ships inside this package at the path given.
Tesseract.js
Files: lib/tesseract/tesseract.min.js
lib/tesseract/worker.min.js
Licence: Apache License 2.0
Text: lib/tesseract/LICENSE-tesseract.txt
Upstream: https://github.com/naptha/tesseract.js
Both bundles embed third-party code of their own, and the notices for it are kept beside them exactly as the upstream build emits them:
lib/tesseract/tesseract.min.js.LICENSE.txt
regenerator-runtime - Copyright (c) 2014-present, Facebook, Inc. - MIT
lib/tesseract/worker.min.js.LICENSE.txt
buffer - Feross Aboukhadijeh - MIT
ieee754 - Feross Aboukhadijeh - BSD-3-Clause
regenerator-runtime - Copyright (c) 2014-present, Facebook, Inc. - MIT
zlib.js - 2012 imaya - MIT
Tesseract OCR engine, compiled to WebAssembly
Files: lib/tesseract/tesseract-core-simd-lstm.js
lib/tesseract/tesseract-core-simd-lstm.wasm
Licence: Apache License 2.0
Text: lib/tesseract/LICENSE-tesseract.txt
Upstream: https://github.com/naptha/tesseract.js-core
https://github.com/tesseract-ocr/tesseract
This is a build of the Tesseract OCR engine. It links Leptonica, libpng and libtiff, each under its own permissive licence, and their full texts live with those projects:
Leptonica https://github.com/DanBloomberg/leptonica libpng https://github.com/pnggroup/libpng libtiff https://gitlab.com/libtiff/libtiff
Japanese vertical recognition data
File: lib/tesseract/jpn_vert.traineddata
(Chrome and Firefox packages only - the Opera store accepts no
.traineddata file, so the Opera package leaves it out)
Licence: Apache License 2.0
Text: lib/tesseract/LICENSE-tesseract.txt
Upstream: https://github.com/tesseract-ocr/tessdata
Recognition data for any other language is never bundled, and in the Opera package that includes Japanese. When one is needed it loads from https://tessdata.projectnaptha.com and is cached on the device, under the same licence.
fflate
File: lib/fflate.min.js Licence: MIT - Copyright (c) 2023 Arjun Barrett Text: lib/LICENSE-fflate.txt Upstream: https://github.com/101arrowz/fflate
Used to read the .zip files of dictionaries and subtitles that the user imports or downloads.
Kanji usage frequency data
Where: the KANJI_FREQ table inside lib/ocr-engine.js Licence: Creative Commons Attribution 4.0 International Text: lib/tesseract/LICENSE-kanji-frequency.txt Upstream: https://github.com/scriptin/kanji-frequency
Derived, with changes, from Dmitry Shpika's datasets. The changes are described in the licence file named above.
Brand marks
Files: icons/youtube.png, icons/twitch.png, icons/instagram.png,
icons/discord.png, icons/patreon.png, icons/kofi.png,
icons/github.png
These are the trade marks of their respective owners. They are shipped unmodified and used for one purpose: as the icons of the links that open Kitsumi's own pages on those services. No affiliation with, sponsorship by, or endorsement from any of them is claimed or implied.
Note that this is about trade marks, not about who drew the file. These icons depict other companies' marks whoever made the image, which is why they are listed here.
Everything else in icons/ - the app icons, the fox artwork, the settings icon and the opening video - is the author's own work and is covered by section 1 of LICENSE, not by anything here. If you believe a file in this package is yours, write to the address in LICENSE and it will be credited or removed.
Sounds
sounds/AppLaunch.mp3 is the author's own work and is covered by section 1 of LICENSE, not by anything here.
sounds/known_add.mp3 and sounds/known_remove.mp3 were synthesised for Kitsumi from plain sine waves by a script (no samples, no third-party audio); they are part of the project and covered by section 1 of LICENSE as well.
The other five clips came from Freesound (https://freesound.org/) and Pixabay (https://pixabay.com/). Each was transcoded to MP3 for this package. The audio itself is unaltered - no trimming, no mixing, no effects. Under section 2(a)(4) of the Creative Commons 4.0 licences, a change of format alone never produces adapted material, so nothing here is an adaptation of anybody's work.
Pixabay content is covered by the Pixabay Content License (https://pixabay.com/service/license-summary/), which permits use in a commercial product and asks for no attribution.
Freesound has no single licence: each upload carries the one its author chose. Nothing here is under a NonCommercial licence. The clips used are:
REQUIRES ATTRIBUTION
"Peanut drop 9.wav" by DarkOrochi
Sound: https://freesound.org/people/DarkOrochi/sounds/322246/
Author: https://freesound.org/people/DarkOrochi/
Licence: Creative Commons Attribution 4.0
https://creativecommons.org/licenses/by/4.0/
NO ATTRIBUTION REQUIRED - Creative Commons 0, public domain https://creativecommons.org/publicdomain/zero/1.0/ Listed anyway, because credit costs nothing and these people gave their work away.
"anime cat girl.wav" by KurireeVA
Sound: https://freesound.org/people/KurireeVA/sounds/476588/
Author: https://freesound.org/people/KurireeVA/
"effect_notify.wav" by ricemaster
Sound: https://freesound.org/people/ricemaster/sounds/278142/
Author: https://freesound.org/people/ricemaster/
"Success notification" by _lucy
Sound: https://freesound.org/people/_lucy/sounds/780010/
Author: https://freesound.org/people/_lucy/
"Pop 4" by quatricise
Sound: https://freesound.org/people/quatricise/sounds/789793/
Author: https://freesound.org/people/quatricise/
Dictionaries offered in the catalog
These are NOT part of this package. Kitsumi hosts none of them and ships none of them. The catalog is a list of links; a file is fetched from a third-party host only when the user asks for it, and it then lives in that user's own browser. Each keeps whatever licence its publisher gave it, and the obligation to honour that licence is the publisher's and the user's, not Kitsumi's.
They are listed here because a person deciding whether to install one should be able to see the terms, and because Kitsumi states the same licence on the row itself, in the catalog, before the download button.
JMdict, JMnedict, KANJIDIC CC BY-SA 4.0
Electronic Dictionary Research and Development Group
https://www.edrdg.org/ build: yomidevs/jmdict-yomitan
Jitendex CC BY-SA 4.0
https://jitendex.org/
Wiktionary builds ("wty") CC BY-SA 4.0
huggingface.co/datasets/daxida/wty-release
Wikipedia builds CC BY-SA
github.com/MarvNC/wikipedia-yomitan
CC-CEDICT CC BY-SA 3.0
https://cc-cedict.org/
KRDICT, the Korean Learners' CC BY-SA 2.0 KR
Dictionary of the National
Institute of Korean Language
build: github.com/Lyroxide/yomitan-ko-dic
Two rows in that same Korean build, the STDICT monolingual dictionary and the CC100 frequency list, state no licence at their source. Kitsumi says so on the row rather than implying terms nobody has stated.
Anything a user imports from their own files is entirely their own business and never appears here.
Online dictionary services
Off until the user switches online lookups on. Each lookup sends the single hovered word to the service; nothing is stored by Kitsumi, and the answer is shown with the service's name on it. Kitsumi is not affiliated with any of them. The data behind them, where the service states it:
Jisho jisho.org
JMdict and KANJIDIC by the EDRDG, CC BY-SA 4.0
Jotoba jotoba.de (AGPL-3.0 service)
JMdict, JMnedict and KANJIDIC2 by the EDRDG, CC BY-SA 4.0
KRDict Open API krdict.korean.go.kr
Korean Learners' Dictionary, National Institute of Korean Language,
CC BY-SA 2.0 KR. Needs the user's own free key.
Wiktionary *.wiktionary.org (Wikimedia Foundation)
Text by Wiktionary contributors, CC BY-SA 4.0. Kitsumi reads the
Explain-language edition and, for English, German, French and Spanish
words, the translation table on the word's own edition.
Free Dictionary API freedictionaryapi.com
Wiktionary data via Wiktextract (kaikki.org), CC BY-SA 4.0
Dictionary API dictionaryapi.dev (GPL-3.0 service)
Wiktionary data, CC BY-SA
MyMemory mymemory.translated.net (Translated srl)
Translation memory and machine translation, free anonymous tier.