{{Short description|Online project collecting example sentences}} {{Use dmy dates|date=November 2014}} {{Infobox website | name = Tatoeba | logo = Tatoeba Logo.svg | logo_size = 125px | logo_caption = The logo of Tatoeba | screenshot = Tatoeba.org English screenshot 2021.png | screenshot_size = 250px | collapsible = 1 | url = {{URL|https://tatoeba.org/}} | commercial = No | type = Online parallel corpora | registration = Optional | language_count = 59 languages of the interface; content in 429 (April 2026) | country_of_origin = France | owner = Association Tatoeba | founder = Trang Ho | key_people = Allan Simon | content_license = CC BY (some sentences under CC0), audio varies | launch_date = 2006 | current_status = Online }}

'''Tatoeba''' is a free collection of example sentences with translations geared towards foreign language learners. It is available in more than 400 languages. Its name comes from the Japanese phrase {{nihongo||例えば|tatoeba}}, meaning 'for example'. It is written and maintained by a community of volunteers through a model of open collaboration. Individual contributors are known as "Tatoebans". It is run by Association Tatoeba, a French non-profit organization funded through donations.

== History and development ==

In 2006, Trang Ho was frustrated that unlike some of their Japanese counterparts, German bilingual dictionaries didn't feature full-text search of usage examples with translations.<ref>{{Cite web |last=Trang |title=The story of Tatoeba |url=http://blog.tatoeba.org/2013/05/the-story-of-tatoeba.html |access-date=2022-11-08}}</ref> It led her to imagine her ideal dictionary<ref>{{Cite web |title=Trang's ideal dictionary.pdf |url=https://drive.google.com/file/d/0ByYR_JAvpNe7SVlxWXRSbXN4cnM/view?resourcekey=0-rCZSKBddbEMHZSbWC6Cs-g&usp=embed_facebook |access-date=2022-11-08 |website=Google Docs}}</ref> and to build a prototype hosted on SourceForge under the name "multilangdict."<ref>{{cite web |title=Trang's dictionary project |url=https://sourceforge.net/projects/multilangdict/ |work=sourceforge.net|date=10 April 2013 }}</ref> The main focus was already the crowdsourcing of translated sentences: "A Wikipedia type of thing, except people add sentences, not articles."

Alongside her studies at the University of Technology of Compiègne, Trang Ho gradually improved her website with a few classmates. She rebuilt the project from scratch twice and rebranded it as Tatoeba. In September 2007, about 150,000 English-Japanese sentence pairs from the Tanaka Corpus — a public-domain compilation released in 2001 by Hyogo University professor Yasuhito Tanaka and maintained by Jim Breen and Paul Blay — were imported into the Tatoeba Corpus.<ref>{{cite web |date=3 February 2011 |title=Tanaka Corpus |url=http://www.edrdg.org/wiki/index.php/Tanaka_Corpus |accessdate=20 March 2011 |work=EDRDG Wiki |publisher=Electronic Dictionary Research and Development Group}}</ref> In December 2008, Trang Ho released the first version of the current codebase built around a more flexible data model.<ref>{{Citation |title=Tatoeba Stream #3 - Going back in time | date=29 November 2021 |url=https://www.youtube.com/watch?v=b0DCcx_4IrU |language=en |access-date=2022-11-08}}</ref> The following month, the website moved to the tatoeba.org domain.<ref>{{Cite web |last=Trang |title=New address : tatoeba.org |url=http://blog.tatoeba.org/2009/01/new-address-tatoebaorg.html |access-date=2022-11-08}}</ref>

Over the 2009-2010 academic year, Allan Simon — then a student at SUPINFO — became a core developer of Tatoeba. Together with Trang Ho and other young developers, they made Tatoeba more social: sentence lists, user profiles, private messaging, and Facebook-inspired Wall. They also introduced significant features like sentence linking, tagging, and "translation of translation" search. In November 2010, Tatoeba passed the 600,000 sentences mark. Within a year, the number of sentences added daily had increased almost 50-fold.<ref>{{Cite web |last=Trang |title=Some stats |url=http://blog.tatoeba.org/2010/10/some-stats.html |access-date=2022-11-08}}</ref>

Between 2014 and 2016, a new team of developers formed around Trang Ho.<ref>{{Cite web |last=AlanF |title=Update on development |url=http://blog.tatoeba.org/2014/02/update-on-development.html |access-date=2022-11-08}}</ref> They mentored students at the Google Summer of Code 2014<ref name=":1">{{Cite web |title=Google Summer of Code 2014 Organization Association Tatoeba |url=https://www.google-melange.com/archive/gsoc/2014/orgs/tatoeba |accessdate=26 September 2022 |website=www.google-melange.com}}</ref> and added features to improve corpus quality.

Over the 2018-2020 period, support from the Mozilla Foundation as part of the Common Voice project allowed Tatoeba to make its platform more open and user-friendly.<ref name=":2">{{Cite web |title=MOSS award for Tatoeba |url=http://blog.tatoeba.org/2018/05/moss-award-for-tatoeba.html |accessdate=26 September 2022}}</ref><ref name=":3">{{Cite web |title=A second MOSS award |url=http://blog.tatoeba.org/2019/08/a-second-moss-award.html |accessdate=26 September 2022}}</ref>

==Openness==

{{Historical populations |title = Active Tatoeba editors |pop_name = Editors |year_name = Year |source = [//downloads.tatoeba.org/exports/contributions.tar.bz2 Tatoeba contributions] |2010 |1399 |2011 |1989 |2012 |2322 |2013 |2377 |2014 |2248 |2015 |2506 |2016 |2085 |2017 |1481 |2018 |1583 |2019 |1420 |2020 |1735 |2021 |1540 |2022 |1377 |2023 |1335 |2024 |1208 |2025 |1302 }}

=== Use ===

Users can search for words to retrieve sentences that use them. Results can be filtered by language, number of words, tag, and other criteria.<ref>{{Cite web |title=Advanced search - Tatoeba |url=https://tatoeba.org/en/sentences/advanced_search |access-date=2023-11-21 |website=tatoeba.org}}</ref>

Each sentence is displayed next to its translations and "translations of translations". A comment section facilitates feedback and corrections.

Registered users can build downloadable lists of sentences, which are private, public or collaborative.

=== Contribution ===

Tatoebans are encouraged to contribute in their strongest language.<ref>{{Cite web |title=Quick Start Guide |url=http://en.wiki.tatoeba.org/articles/show/quick-start}}</ref> They can add original sentences and translate existing ones. They can proofread or comment on other users' sentences, and "adopt" sentences without an owner. Advanced contributors are also allowed to tag, link, and unlink sentences.

When the owner of a sentence does not respond to a correction request, only a corpus maintainer has the power to update or delete the sentence.

== Governance == As founder of Tatoeba, Trang Ho has long been the project's BDFL.

In 2011, she set up a nonprofit organization to oversee the project.

In 2022, she decided to step aside in favor of a small group of experienced Tatoebans.<ref>{{Cite web |title=Thread #38883 - Tatoeba |url=https://tatoeba.org/en/wall/show_message/38883#!#message_38883 |access-date=2023-11-21 |website=tatoeba.org}}</ref>

== Languages == thumb|A simplified diagram of Tatoeba's underlying data structure.|335x335pxAs of April 2026, the Tatoeba Corpus has over 13,400,000 sentences in 429&nbsp;languages. 69&nbsp;of these languages have 10,000&nbsp;or more sentences. Over 1&nbsp;million sentences have audio recordings.<ref name=":4">{{Cite web |title=Number of sentences per language - Tatoeba |url=https://tatoeba.org/en/stats/sentences_by_language |access-date=2025-02-22 |website=tatoeba.org}}</ref>

The sentences are interrelated within a graph that has more than 25,900,000&nbsp;links. 276&nbsp;language pairs have over 10,000&nbsp;translated sentences.<ref name=":0" />

{{Bar chart | title = The 20 languages with the most links (as of February 2026)<ref>[https://tatoeba.org/en/downloads Tatoeba weekly exports]</ref> | label_type = Language | data_type = Number of links | width_units = em | data_max = 7500000 | label1 = English | data1 = 7,314,995 | label2 = Russian | data2 = 2,261,268 | label3 = French | data3 = 2,197,345 | label4 = German | data4 = 2,041,759 | label5 = Esperanto | data5 = 2,028,872 | label6 = Italian | data6 = 1,169,344 | label7 = Spanish | data7 = 1,136,238 | label8 = Turkish | data8 = 929,615 | label9 = Dutch | data9 = 704,617 | label10 = Portuguese | data10 = 671,970 | label11 = Japanese | data11 = 609,652 | label12 = Hungarian | data12 = 479,026 | label13 = Ukrainian | data13 = 444,988 | label14 = Kabyle | data14 = 396,084 | label15 = Hebrew | data15 = 296,191 | label16 = Finnish | data16 = 242,226 | label17 = Polish | data17 = 226,637 | label18 = Mandarin Chinese | data18 = 221,810 | label19 = Danish | data19 = 192,704 | label20 = Lithuanian | data20 = 159,629 |bar_width=35 }}

== Operation ==

Tatoeba received a grant from Mozilla Drumbeat in December 2010.<ref>{{cite web |last=Ho |first=Trang |date=17 January 2011 |title=Grant from Mozilla Drumbeat |url=http://blog.tatoeba.org/2011/01/grant-from-mozilla-drumbeat.html |accessdate=20 March 2011 |work=Tatoeba Project Blog}}</ref><ref>{{cite web |last=Moltke |first=Henrik |date=30 December 2010 |title=Best Drumbeat Projects: Tatoeba – a free and open database of sentences |url=http://yoyodyne.cc/tatoeba/ |url-status=dead |archiveurl=https://web.archive.org/web/20110102095352/http://yoyodyne.cc/tatoeba/ |archivedate=2 January 2011 |accessdate=20 March 2011 |work=Yoyodyne.cc |quote=...the Mozilla Foundation wants to encourage and help the Tatoeba project by giving it a USD 2.5K Mozilla Drumbeat Grant. |df=dmy-all}}</ref>

Some work on the Tatoeba infrastructure was sponsored by Google Summer of Code, 2014 edition.<ref name=":1" />

Since 2014, Tatoeba has been supported by donations.<ref>{{cite web |title=Donations |url=https://en.wiki.tatoeba.org/articles/show/donations |website=en.wiki.tatoeba.org}}</ref>

In May 2018 they received a $25,000 Mozilla Open Source Support (MOSS) program grant.<ref name=":2" />

In August 2019 they received a $15,000 Mozilla Open Source Support (MOSS) program grant.<ref name=":3" />

== Access to content ==

=== Licensing === By default, the sentences of the Tatoeba Corpus are published under a CC BY license,<ref>{{cite web |url=https://tatoeba.org/eng/terms-of-use |title=Terms of use |work=Tatoeba.org |accessdate=20 March 2011}}</ref> freeing it for academic and other use. Users can also contribute sentences under CC0, though translations of those sentences currently can't share the same license.<ref>{{Cite web|title=How to contribute under CC0|url=https://en.wiki.tatoeba.org/articles/show/cc0-contributions|access-date=2021-10-25|website=en.wiki.tatoeba.org}}</ref>

Audio recordings of the sentences use the speaker's choice of license, such as CC BY, CC BY-SA, CC BY-NC, or no public license at all.<ref>{{Cite web|title=All public lists containing "audio" (140) - Tatoeba|url=https://tatoeba.org/en/sentences_lists/index/audio|access-date=2021-10-25|website=tatoeba.org}}</ref>

=== Offline use === Visitors can download tab-delimited sentence pairs ready for import into Anki and similar Spaced Repetition Software at the Tatoeba website.<ref name=":0">{{Cite web |title=Download sentences - Tatoeba |url=https://tatoeba.org/en/downloads |access-date=2025-02-22 |website=tatoeba.org}}</ref>

=== Software development tools === An unstable API is available for software developers.<ref>{{Cite web |title=Tatoeba API |url=https://api.dev.tatoeba.org/unstable#?route=overview |access-date=2023-11-21 |website=api.dev.tatoeba.org}}</ref>

==Related projects==

=== Second-language acquisition === Tatoeba sentences can be used to build lexicographic references for language learners. The JMdict Japanese-English dictionary selects its example sentences from the Tatoeba Corpus.<ref>{{Cite web |title=WWWJDIC - INFORMATION |url=https://www.edrdg.org/wwwjdic/wwwjdicinf.html#examp_tag |access-date=2022-11-13 |website=www.edrdg.org}}</ref> OpenRussian is a free Russian dictionary built primarily from the content of Wiktionary and Tatoeba.<ref>{{Cite web |title=About OpenRussian |url=https://en.openrussian.org/about |access-date=2022-11-16 |website=en.openrussian.org}}</ref> GoodExample tries to automatically extract a diverse set of high-quality example sentences from the English Tatoeba Corpus.<ref>{{Cite web |title=Legal considerations - GoodExample |url=https://www.goodexample.is/legal.html |access-date=2022-12-06 |website=www.goodexample.is}}</ref>

Tatoeba datasets can power incidental learning experiences that blend the acquisition of a foreign language with the user's everyday activities like web browsing or book reading.<ref>{{Cite book |last=Winiwarter |first=Werner |title=Proceedings of the 17th International Conference on Information Integration and Web-based Applications & Services |chapter=JILL |date=2015-12-11 |chapter-url=https://doi.org/10.1145/2837185.2837191 |series=iiWAS '15 |location=New York, NY, USA |publisher=Association for Computing Machinery |pages=1–9 |doi=10.1145/2837185.2837191 |isbn=978-1-4503-3491-4|s2cid=2130581 }}</ref><ref>{{Cite web |title=Lisons! |url=https://fauu.github.io/lisons/ |access-date=2022-12-02 |website=fauu.github.io}}</ref> A team at MIT Media Lab used example sentences from Tatoeba in WordSense, a mixed reality platform that enables "serendipitous language learning in the wild."<ref>{{Cite book |last1=Vazquez |first1=Christian David |last2=Nyati |first2=Afika Ayanda |last3=Luh |first3=Alexander |last4=Fu |first4=Megan |last5=Aikawa |first5=Takako |last6=Maes |first6=Pattie |title=Proceedings of the 2017 CHI Conference Extended Abstracts on Human Factors in Computing Systems |chapter=Serendipitous Language Learning in Mixed Reality |date=2017-05-06 |chapter-url=https://doi.org/10.1145/3027063.3053098 |series=CHI EA '17 |location=New York, NY, USA |publisher=Association for Computing Machinery |pages=2172–2179 |doi=10.1145/3027063.3053098 |isbn=978-1-4503-4656-6|s2cid=1557887 }}</ref> More recently, Japanese researchers implemented a Tatoeba search feature in an integrated writing assistance environment.<ref>Masato Hagiwara, Takumi Ito, Tatsuki Kuribayashi, Jun Suzuki, and Kentaro Inui. 2019. [https://arxiv.org/pdf/1909.02621.pdf TEASPN: Framework and Protocol for Integrated Writing Assistance Environments.] In ''Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP): System Demonstrations'', pages 229–234, Hong Kong, China. Association for Computational Linguistics.</ref>

Although the sentences in the Tatoeba&nbsp;Corpus are not all authentic, they are sometimes used to build data-driven learning applications. BES (Basic English Sentence) Search is a non-commercial tool for finding beginner-level English sentences for use in teaching materials.<ref>{{Cite web |title=BES Search |url=https://bessearch.ddl-study.org/ |access-date=2023-06-14 |website=bessearch.ddl-study.org}}</ref> It has over 1 million sentences, most of them from Tatoeba.<ref>NISHIGAKI, C., & AKASEGAWA, S. [https://jaecs.com/jnl/ECS30/ECS30_131-149.pdf Secondary School Students: What We Can Do to Nurture Autonomous Corpus Users?.]</ref> Reverso uses Tatoeba parallel corpora in its commercial bilingual concordancer.<ref>{{Cite web |title=Reverso Context {{!}} Legal considerations about corpora used in the contextual dictionary |url=https://context.reverso.net/translation/legal |access-date=2022-12-02 |website=context.reverso.net |language=en}}</ref>

Example sentences are also used as a base for exercises. Charles Kelly and Paul Raine, both EFL teachers in Japan, have developed language learning activities based on sentences curated from the Tatoeba&nbsp;Corpus.<ref>Kelly, Charles (2012). [https://core.ac.uk/download/pdf/227969104.pdf "タトエバ・プロジェクト・コーパスを使った www. ManyThings. org の語学学習教材"] (PDF), 愛知工業大学研究報告 (47), 77-84.</ref><ref>{{Cite journal |last=Raine |first=Paul |date=2018 |title=Building Sentences with Web 2.0 and the Tatoeba Database |url=http://www.issues.accentsasia.org/issues/10-2/Raine.pdf |journal=Accents Asia}}</ref> Clozemaster is a language self-study program that generates gamified cloze tests from Tatoeba sentence pairs.<ref>{{Cite web |date=2017-10-17 |title=What is a Cloze Test? Cloze Deletion Tests and Language Learning |url=https://www.clozemaster.com/blog/cloze-test/ |access-date= |website=Clozemaster Blog |language=en-US}}</ref> Some Anki users share flashcards that were created using Tatoeba.<ref>{{Cite web |title=Tatoeba - AnkiWeb |url=https://ankiweb.net/shared/decks/tatoeba |access-date=2022-12-02 |website=ankiweb.net}}</ref>

=== Regional or minority languages === Some language digital activists contribute to open collaborative projects like Tatoeba, Wikipedia, and Common Voice to promote their minority language in digital spaces.<ref>{{Cite web |date=2022-06-28 |title=Rising Voices - Meet Prasanta Hembram, a Santali language digital activist from India |url=https://rising.globalvoices.org/blog/2022/06/28/meet-prasanta-hembram-a-santali-language-digital-activist-from-india/ |access-date=2022-11-15 |website=Rising Voices |language=en}}</ref> Regional languages like Kabyle, Catalan, or Basque can register more than a hundred members on Tatoeba.<ref>{{Cite web |title=Languages of members - Tatoeba |url=https://tatoeba.org/en/stats/users_languages |access-date=2022-11-15 |website=tatoeba.org}}</ref>

=== Constructed languages === Selected content from Tatoeba in Esperanto is available in the multilingual DVD ''Esperanto Elektronike'' published by E@I.<ref>{{Cite web |date=2017-10-13 |title=Esperanto Elektronike {{!}} E@I |url=https://ikso.net/services/esperanto-elektronike-2/ |access-date=2022-11-01 |language=en-GB}}</ref> As of November 2022, Esperanto is Tatoeba's fifth pivot language, with over 330,000 sentences translated into at least two languages.<ref name=":0" /> Other constructed languages like Toki Pona, Interlingua, Klingon, Lojban, and Ido also have a significant footprint.<ref name=":4" />

=== Language technology === thumb|Research articles about machine translation that mention Tatoeba<ref>{{Cite web |title=Google Scholar |url=https://scholar.google.com/scholar?q=%22machine%20translation%22%20Tatoeba |access-date=2022-11-13 |website=scholar.google.com}}</ref> From 2008 to 2011, Francis Bond used the Tatoeba Corpus for his research on the Japanese language.<ref>Francis Bond, 栗林 孝行 [Takayuki Kuribayashi], 橋本 力 [Hashimoto Chikara] (2008) HPSGに基づくフリーな日本語ツリー バンクの構築 [A free Japanese Treebank based on HPSG]. In 14th Annual Meeting of The Association for Natural Language Processing, Tokyo.</ref><ref>Eric Nichols, Francis Bond, Darren Scott Appling and Yuji Matsumoto (2010) Paraphrasing Training Data for Statistical Machine Translation. Journal of Natural Language Processing, 17(3), pages 101–122.</ref>

Since 2013, Jörg Tiedemann has been spreading Tatoeba parallel corpora more widely in the machine translation community by sharing them on the OPUS repository and organizing the "Tatoeba&nbsp;Translation&nbsp;Challenge".<ref>{{Cite web |date=2013-07-30 |title=OPUS - an open source parallel corpus |url=http://opus.lingfil.uu.se/ |access-date=2022-11-13 |archive-url=https://web.archive.org/web/20130730085147/http://opus.lingfil.uu.se/ |archive-date=30 July 2013 }}</ref><ref>{{Cite arXiv |last=Tiedemann |first=Jörg |date=2020-10-13 |title=The Tatoeba Translation Challenge -- Realistic Data Sets for Low Resource and Multilingual MT |class=cs.CL |eprint=2010.06354}}</ref> With the rise of deep learning, researchers increasingly use Tatoeba's data sets to train and evaluate their massively multilingual models in tasks like machine translation,<ref>{{Cite arXiv |last1=NLLB Team |last2=Costa-jussà |first2=Marta R. |last3=Cross |first3=James |last4=Çelebi |first4=Onur |last5=Elbayad |first5=Maha |last6=Heafield |first6=Kenneth |last7=Heffernan |first7=Kevin |last8=Kalbassi |first8=Elahe |last9=Lam |first9=Janice |last10=Licht |first10=Daniel |last11=Maillard |first11=Jean |last12=Sun |first12=Anna |last13=Wang |first13=Skyler |last14=Wenzek |first14=Guillaume |last15=Youngblood |first15=Al |date=2022-08-25 |title=No Language Left Behind: Scaling Human-Centered Machine Translation |class=cs.CL |eprint=2207.04672}}</ref> language identification,<ref>{{Cite web |title=Language identification · fastText |url=https://fasttext.cc/index.html |access-date=2022-11-16 |website=fasttext.cc |language=en}}</ref> semantic search,<ref>{{Cite arXiv |last1=Hu |first1=Junjie |last2=Ruder |first2=Sebastian |last3=Siddhant |first3=Aditya |last4=Neubig |first4=Graham |last5=Firat |first5=Orhan |last6=Johnson |first6=Melvin |date=2020-09-04 |title=XTREME: A Massively Multilingual Multi-task Benchmark for Evaluating Cross-lingual Generalization |class=cs.CL |eprint=2003.11080}}</ref> and speech recognition.<ref>{{Cite arXiv |last1=Wang |first1=Changhan |last2=Pino |first2=Juan |last3=Wu |first3=Anne |last4=Gu |first4=Jiatao |date=2020-06-09 |title=CoVoST: A Diverse Multilingual Speech-To-Text Translation Corpus |class=cs.CL |eprint=2002.01320}}</ref>

==See also== {{Portal|Language|Linguistics}} * Phrase book * Parallel text * Common Voice * Lingua Libre * Wiktionary

==References== <references />

==External links== {{Commonscat}} * {{Official site|https://tatoeba.org}} * [https://www.youtube.com/watch?v=lHotYMpKbr8 Video of Trang Ho introducing Tatoeba at MozFest 2019] * [https://tatoeba.j-langtools.com/allstats/ Tatoeba's statistics] * [https://github.com/Helsinki-NLP/Tatoeba-Challenge Tatoeba Translation Challenge]

{{Corpus linguistics}}

Category:Computational linguistics Category:Corpora Category:Creative Commons-licensed websites Category:Free-content websites Category:French educational websites Category:Language learning software Category:Natural language processing Category:Open data Category:Open educational resources Category:Social networking language-learning websites Category:Free language learning software