How the converter decides the reading
A character-by-character lookup fails on the most common words in the language. The fix is to read phrases, not characters.
Words first
Your text is segmented into words before any lookup. 银行 is looked up as a whole word, so it gets yínháng — not the yíngxíng a character-by-character converter produces.
Tone marks, not numbers
Output uses real tone marks: mā má mǎ mà. Hover any character to see its reading and meaning from the site's character library.
Nothing is uploaded
Everything runs in your browser. Long texts, private messages, exam drafts — paste freely; nothing is sent or stored.
Made for names too
Generated a Chinese name? Paste it here to confirm the reading, then follow each character to its stroke-order animation and practice sheet.
Common questions
Why do some characters change their pinyin?
Many hanzi are polyphonic — the same character carries two or more readings depending on the word it sits in. 行 is xíng in 行走 but háng in 银行; 长 is zhǎng in 长大 but cháng in 长度. This converter segments text into words first and looks up the whole word.
Is my text uploaded anywhere?
No. Segmentation and conversion run entirely in your browser. Nothing you paste is sent, stored or logged.
Does it work for names?
Yes — paste the name you generated on the home page and every character gets its reading with tone marks, which is exactly what you hand to a teacher or write on a name card.
What does the question mark mean?
A character outside the built-in table. Rather than guessing a reading and being confidently wrong, the converter marks it honestly. Rarer characters will gain coverage as the library grows.