Wraps PDFMathTranslate-next to translate academic PDFs while preserving mathematical formulas and layout. Exposes three main tools: translate_pdf for converting documents between Chinese and English, pdf_get for inspecting file metadata like page count and size, and warm_up_translator to preload fonts and models before your first translation. The server handles relative and absolute paths, reports progress during translation, and generates both dual-language and target-language-only outputs. First run downloads about 50MB of assets, then translations run in under a minute for typical papers. Reach for this when you need to translate technical documents with equations without mangling the formatting or spending time configuring translation pipelines yourself.
claude mcp add --transport stdio huoshuiai42-huoshui-pdf-translator -- uvx huoshui-pdf-translator