Our written sentences and translations can be used under the Creative Commons Attribution 2.0 license (CC-BY 2.0).More information
If you love this content, please consider a donation.
Note
If you simply want sentences that you can use to learn a language, check out the sentence lists. You can build your own, or view the ones that others have created. The lists can be downloaded and printed.
General information about the files
The files provided here are updated every Saturday at 9 a.m. (GMT).
Many of the Japanese and English sentences are from the Tanaka Corpus, which belongs to the public domain.
Creative commons
These files are released under CC-BY.
Licenses covering audio
The license covering an audio file is chosen by the contributor, and is indicated on the page that lists the audio files that he or she has contributed.
Questions?
If you have questions or requests, feel free to contact us. In general, we answer quickly.
Downloads
Sentences
- Download
- 1. http://downloads.tatoeba.org/exports/sentences.tar.bz2
- 2. http://downloads.tatoeba.org/exports/sentences_detailed.tar.bz2
- Fields and structure
- 1. Sentence id [tab] Lang [tab] Text
- 2. Sentence id [tab] Lang [tab] Text [tab] Username [tab] Date added [tab] Date last modified
- File description
- Contains all the sentences. Each sentence is associated with a unique id and an ISO 639-3 language code.
Links
- Download
- http://downloads.tatoeba.org/exports/links.tar.bz2
- Fields and structure
- Sentence id [tab] Translation id
- File description
- Contains the links between the sentences. 1 [tab] 77 means that sentence #77 is the translation of sentence #1. The reciprocal link is also present, so the file will also contain a line that says 77 [tab] 1 .
Tags
- Download
- http://downloads.tatoeba.org/exports/tags.tar.bz2
- Fields and structure
- Sentence id [tab] Tag name
- File description
- Contains the list of tags associated with each sentence. 381279 [tab] proverb means that sentence #381279 has been assigned the "proverb" tag.
Lists
- Download
- http://downloads.tatoeba.org/exports/user_lists.tar.bz2
- Fields and structure
- List id [tab] Username [tab] Date created [tab] Date last modified [tab] List name [tab] Editable by
- File description
- Contains the list of sentence lists.
Sentences in lists
- Download
- http://downloads.tatoeba.org/exports/sentences_in_lists.tar.bz2
- Fields and structure
- List id [tab] Sentence id
- File description
- Indicates the sentences that are contained by any lists. 13 [tab] 381279 means that sentence #381279 is contained by the list that has an id of 13.
Japanese indices
- Download
- http://downloads.tatoeba.org/exports/jpn_indices.tar.bz2
- Fields and structure
- Sentence id [tab] Meaning id [tab] Text
- File description
- Contains the equivalent of the "B lines" in the Tanaka Corpus file distributed by Jim Breen. See this page for the format. Each entry is associated with a pair of Japanese/English sentences. Sentence id refers to the id of the Japanese sentence. Meaning id refers to the id of the English sentence.
Sentences with audio
- Download
- http://downloads.tatoeba.org/exports/sentences_with_audio.tar.bz2
- Fields and structure
- Sentence id [tab] Username [tab] License [tab] Attribution URL
- File description
- Contains the ids of the sentences, in all languages, for which audio is available. Other fields indicate who recorded the audio, its license and a URL to attribute the author. If the license field is empty, you may not reuse the audio outside the Tatoeba project.
User skill level per language
- Download
- http://downloads.tatoeba.org/exports/user_languages.tar.bz2
- Fields and structure
- Lang [tab] Skill level [tab] Username [tab] Details
- File description
- Indicates the self-reported skill levels of members in individual languages.
Users' sentence ratings
- Download
- http://downloads.tatoeba.org/exports/users_sentences.csv
- Fields and structure
- Username [tab] Lang [tab] Sentence id [tab] Rating [tab] Date added [tab] Date last modified
- File description
- Contains sentences rated by users. The value of the rating can be -1 (sentence not OK), 0 (undecided or unsure), or 1 (sentence OK). Warning: this data is still experimental.