deusyu/translate-book
VerifiedClaude Code skill for parallel translation of full books in multiple formats.
What is deusyu/translate-book?
deusyu/translate-book is an open-source Claude Code skill that manages end-to-end book translation. It accepts common ebook formats and produces translated versions in HTML, DOCX, EPUB, or PDF while preserving structure and allowing selective re-runs.
The workflow converts input files to markdown chunks tracked by a manifest, then assigns each chunk to a dedicated subagent for isolated translation. Neighbor excerpts provide limited context, and hash validation ensures output consistency before merging results.
It suits translators, publishers, and researchers who need reliable multi-language book conversions without manual chunking or repeated full-context sessions in a single Claude interaction.
Capabilities
What you can build with deusyu/translate-book
Academic paper localization
Convert research PDFs into another language while keeping chapter structure and references intact for review by international teams.
Fiction series adaptation
Translate EPUB novels chapter by chapter with parallel agents and resume from any point if glossary updates require re-processing.
Technical manual distribution
Produce DOCX and PDF versions of documentation in multiple languages with consistent terminology enforced across batches.
Install deusyu/translate-book
npx skills add deusyu/translate-book -a claude-code -gnpx skills add deusyu/translate-book -a claude-code -g- 1Install via npx with the skills add command targeting Claude Code.
- 2Ensure Calibre, Pandoc, and required Python packages are available on the system.
- 3Run the translate command inside Claude Code pointing to a PDF, DOCX, or EPUB file and target language.
- 4Monitor progress through chunked subagent activity and manifest validation.
- 5Locate all generated files inside the book-specific temp folder for review or further editing.
Works with
deusyu/translate-book: pros & cons
Pros
- +Parallel subagents reduce total translation time and avoid context overflow.
- +Manifest-based hashing and resume support prevent wasted work on partial failures.
- +Unified pipeline produces multiple output formats from one run.
- +Neighbor context helps maintain pronoun and entity consistency across chunks.
Cons
- –Requires separate installation of Calibre and Pandoc before use.
- –Performance depends on Claude API rate limits and available concurrency.
- –Contribution process favors detailed issues over direct pull requests.
Frequently asked questions
The skill currently handles Chinese, English, Japanese, Korean, French, German, and Spanish, with easy extension for others.
User reviews
Verified reviews from the community shape this listing's rating.
Loading reviews…