Free online tool — no sign-up
The other direction is hard in a different way. What goes missing is not vocabulary — it is the entire layer of social information carried by anh, chị, em, cô, chú, bác.
The first 15 pages of text are sent to our server to be translated. The file itself stays on your device.
Vietnamese addresses people by relationship. Every time a character says em không biết, a Vietnamese reader learns roughly how old these two people are relative to each other and how close they stand. Render it as I don't know and the sentence survives intact while all of that quietly disappears.
There is no lossless way to translate it. There is a way to lose it consistently: read the whole book first, work out who everyone is, then decide once for the entire book — where to keep the term, where to substitute a name, where to let context carry it.
The second issue gets less attention. Sentence length. Vietnamese prose runs long, joined by commas, and reads perfectly naturally that way. Carried straight into English it becomes the kind of run-on nobody writes. Because the passage is translated as continuous prose rather than sentence by sentence, those chains get re-cut where English wants them cut.
A DRM-free EPUB. The file never leaves your device; only the text is sent, and only while the job runs.
It records who each character is and how they address one another — exactly the information kinship terms carry and English has nowhere to put.
About 4,500 words, translated as continuous prose so long Vietnamese clause chains can be re-cut for English rhythm.
Kinship terms cannot be translated, only handled. Every choice loses something. The guide only guarantees the whole book loses it the same way, instead of each chapter differently.
Poetry and lục bát will not survive. Metre and internal rhyme are the first things machine translation drops.
Fifteen pages, not the book. The rest stays in Vietnamese in the same file.
The text is sent to a server and discarded when the job ends. The file is never uploaded.
Decided once from the whole book, then held. Sometimes the term is kept, sometimes replaced with a name, sometimes carried by context. There is no perfectly right answer here, only a consistent one.
It will translate it, but expect errors. Without tone marks a single written syllable can be half a dozen different words, and context does not fully resolve that. Use a properly typed source if one exists.
Yes. Personal and place names keep their diacritics, because that is their correct spelling rather than a font problem.
For letting an editor judge the content, yes. For a literary translation to be published, no — a machine keeps the meaning and loses the writing. Use this to open the door, not to walk through it.
All free AI book tools · Free converters that run in your browser
The app translates the entire book with the same engine and puts the Vietnamese beside the English so you can check any line against its source.
Open DualBook