MinerU PDF Translate skill
MinerU PDF Translate turns scientific PDFs into translated PDFs in another language with formulas and figures preserved.
- 100 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- PDF contents go to the MinerU API and the chosen model, which need a token and a key.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/mineru-pdf-translateDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- PDF contents go to the MinerU API and the chosen model, which need a token and a key.
This is third-party code. Review the repository files before installing.
What it does
A Codex skill and a standalone Python script process a folder of PDFs in three steps. MinerU parses the document into Markdown and image references, an OpenAI-compatible model translates the text, and a headless browser prints the result back to PDF. Image references are protected from translation and MathJax renders formulas. The `--ocr-correction auto` mode shows the model crops of the source page to fix recognition errors in formulas and tables. The `--target-language` and `--target-suffix` flags set the output language, with Simplified Chinese as the default.
Who it is for. Researchers and students who read English papers and want the translation as a PDF in the same layout.
Good fit when
- A batch of papers in a folder must be translated with one command.
- A paper is full of formulas and plain Markdown translation breaks them.
- The source PDF was recognized with errors that need fixing from the page image.
- After style changes the PDF must be re-rendered from the translation cache.
Not a fit when
- The document cannot be sent to the external MinerU or model API.
- You need translation without a MinerU token and a model key.
- The source is a scanned book with no structure.
Example request
Translate every PDF in this folder into Russian, keep formulas and figures and check the last appendix pages.Limitations
It needs a MinerU token, a base URL and key of an OpenAI-compatible API, Python 3.10 or newer, Node.js with the `playwright` package and an installed Edge, Chrome or Chromium. The README names key files in Chinese; environment variables are simpler. The author notes that OCR correction is conservative but cannot guarantee semantic accuracy, so complex formulas and appendices should be checked against the original. Output appears only after calls to two external services, so large batches of papers take noticeable time and tokens.
How to disable. Delete the `mineru-pdf-translate` folder from the Codex skills directory. Temporary files live in `.pdf_translate_tmp`, finished PDFs in `translated`, keys in files beside the documents or in environment variables.
FAQ
Where does the result go?
Into a `translated` folder beside the source PDFs. The `--output-dir` flag sets another path.
Can it translate into a language other than Chinese?
Yes. Add `--target-language "Japanese"` and `--target-suffix ja`; language and file suffix are chosen independently.
How do I rebuild a PDF without translating again?
Run the script with `--render-only --keep-temp`; it re-renders the PDF from the cached translated Markdown.
Which keys are needed?
A MinerU token and the base URL and key of an OpenAI-compatible API. Set them through flags, files in the PDF folder or variables such as `MINERU_API_TOKEN`.
What does `--ocr-correction auto` do?
It shows the model crops of the source page with recognized formula and table strings so OCR errors are fixed separately from prose translation.
Discussion
No questions or reviews yet.
Related
SkillsA zero-key, zero-dependency skill and CLI that turns PDF, Office and image files into Markdown via MinerU's cloud.
Official MinerU MCP server: PDFs, Word, PowerPoint and images become Markdown with formulas and tables
MinerU is a CLI with a document library: search and revisit an already parsed report using its local cache.
Translates scientific PDFs while keeping formulas, tables and layout intact, as a CLI, Docker image and MCP server for agents
A skill translates an academic paper PDF into a new PDF whose pages match the original one to one, with formulas in LaTeX.
Five skills turn a research paper PDF into slides, a poster, a project page, a Xiaohongshu post and a WeChat article.
README in short
The README shows a demo and the three pipeline steps, lists features, repository structure and requirements. Two ways to run, key setup, command examples and the main options follow.