Who runs it

Cheng Xu (Xu Cheng), a Singapore-based independent developer, builds and edits the site. Contact: hello@sentencecoach.com.

Data sources

HSK 2.0 word lists and the 2021 standards lists, pinyin and frequency: complete-hsk-vocabulary (MIT licence).
English meanings and measure words: CC-CEDICT, licensed CC BY-SA 4.0. Pages that show these definitions share them under the same licence.
Pinyin for example sentences is generated with pinyin-pro and checked by hand on guide pages.

Open dataset & research citation

We publish the SentenceCoach HSK 1–4 Chinese Sentence Ordering Dataset (v1.0.0, 2,099 segmented 连词成句 exercises with pinyin and English glosses) under CC BY-SA 4.0 for teachers, researchers and NLP developers:

Zenodo archive: DOI 10.5281/zenodo.23951529
GitHub repository: xucheng2024/sentencecoach-open-data
Hugging Face: eatfreshapple/sentencecoach-hsk-sentence-ordering
Kaggle: sentencecoach-hsk-1-4-chinese-sentence-ordering

How the checker works

Your sentence is sent to a large language model (DeepSeek) with instructions written for learner errors: minimal corrections, one plain-English rule per error, no stylistic rewrites. Pinyin is added on our side. Feedback is AI-generated and can be wrong, especially for creative or very advanced writing. Sentences are not stored with your identity; see Privacy.

Editorial standards & classroom use

Grammar and measure-word guides are AI-assisted drafts written and checked by the editor against standard references. Word usage notes are added level by level, and a word page is only indexed once its notes have been reviewed. Teachers and bloggers can embed the sentence checker or print A4 HSK worksheets for free.