Webnovels AI

The auto glossary is the core of the product. The system extracts names, places, and specialized terms from a book, builds a list, and then forces the translation to follow it. You can export the list, edit it in a spreadsheet, and import it back. This fixes the worst failure of plain MTL, the machine translation route most readers still use. If your protagonist is named Wei Ying and the translator keeps calling him Wei Ying in one chapter and Wei Yin in the next, the glossary stops it.

Nobody enjoys building a term list by hand, so the platform generates one for you before you translate. The team warns that extraction is thorough to a fault, meaning your job is trimming rather than adding. The tradeoff is time: the tool parses the entire novel to find candidates, so generation takes a while on long books.

You upload a book and get a book back. The product is built for full-length novels rather than short excerpts, and the team reports that a 200,000 to 300,000 character novel typically finishes within about an hour. That's the practical reason people use it. Chapter by chapter was never the bottleneck for readers, waiting months for an official release was.

Finished translations download as EPUB, PDF, or TXT. EPUB translation is the format most readers pick, and stay available in the account for online reading. EPUB keeps chapter breaks, so the file opens in a normal reader app with a working table of contents. TXT suits anyone feeding the text into another tool. PDF is the least useful for long novels, but it exists.

The engine handles Chinese, Japanese, Korean, and Thai, mostly into English. Each source gets its own page: Chinese novel translation, Korean novel translation, and Japanese novel translation are treated as separate services. Separate pages cover simplified Chinese, traditional Chinese, Japanese, Korean, Vietnamese, Thai, and Indonesian, each treated as its own service with its own sample. The bias is clear: this is an English-first tool for readers of Asian fiction.

Beyond files, the platform translates text inside images, which covers manga and manhwa pages where dialogue is baked into the artwork. A web page translator lets you paste a URL and translate the content directly. Neither feature replaces the book workflow, but they're handy when a story only exists as a serialized web page.

Raw epubs from fan sites often carry junk: stray tags, broken paragraphs, duplicated headers. The platform cleans the text automatically and splits chapters into consistent units before translation. The result reads like a book, not like a scrape.