A sophisticated document parsing tool that converts complex PDFs, Word docs, and PPTs into AI-ready Markdown with full support for formulas and tables.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install mineru
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install mineru using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is MinerU?
MinerU is a professional-grade document parsing solution developed by OpenDataLab, designed to solve the challenge of extracting clean, structured data from unstructured file formats. It specialized in converting PDF, Word, and PPT files into high-fidelity Markdown, ensuring that complex elements like LaTeX formulas and multi-column layouts remain intact. For developers building Openclaw Skills, MinerU serves as a critical bridge for ingesting academic papers and technical reports into LLM-based workflows.
Beyond simple text extraction, MinerU utilizes advanced layout analysis and OCR technologies to recognize table structures and images accurately. This makes it an indispensable tool for researchers and developers who need to automate the digitization of technical documentation without losing the semantic meaning of mathematical notations or tabular data.
MinerU Use Cases
- Automatically converting academic arXiv papers into Markdown for LLM summarization and analysis.
- Ingesting legacy corporate PDF reports into a knowledge base for RAG-based Openclaw Skills.
- Batch processing scanned documents using OCR to create searchable text archives.
- Extracting complex mathematical formulas in LaTeX format for technical documentation updates.
How MinerU Works
- The user provides a document URL or uploads a file to the MinerU API via a POST request.
- The system initiates a layout analysis using models like doclayout_yolo or layoutlmv3 to identify text, tables, and images.
- Specialized OCR and formula recognition engines process the visual elements to convert them into structured text and LaTeX code.
- The parsing status is tracked via a task ID until the processing lifecycle is complete.
- A structured ZIP package is generated containing the full Markdown content, metadata, and extracted image assets for integration into Openclaw Skills.
MinerU Setup
To use MinerU, you must first obtain an API key from the official portal and configure your environment.
# Set your API token in your shell or Openclaw configuration
export MINERU_TOKEN="your_api_key_here"
You can then submit tasks using curl:
curl -X POST "https://mineru.net/api/v4/extract/task" \
-H "Authorization: Bearer $MINERU_TOKEN" \
-H "Content-Type: application/json" \
-d '{
"url": "https://arxiv.org/pdf/2410.17247",
"enable_formula": true,
"enable_table": true
}'
MinerU Data Schema & Taxonomy
Upon completion, the skill provides a structured output directory within a ZIP archive:
| File/Folder | Description |
|---|---|
full.md |
The primary output containing the complete Markdown version of the document. |
content_list.json |
A detailed JSON schema of the extracted content for programmatic processing. |
images/ |
A folder containing all figures, charts, and images extracted from the source. |
layout.json |
Metadata regarding the layout analysis and coordinate data for all identified blocks. |
MinerU Advanced Features
- Vision Language Model (VLM) Support: Switch to the vlm model for maximum accuracy on highly complex layouts.
- Batch Processing: Submit multiple files simultaneously to optimize API quota and throughput for large-scale Openclaw Skills automation.
- Layout Model Selection: Choose between high-speed YOLO-based models or higher-accuracy LM-based models depending on the document complexity.
- Multi-language OCR: Robust support for mixed-language documents, including English and Chinese characters.
SKILL.md
Loading
Related Openclaw Skills
Kitful AI Article Generator
An AI-driven content engine that transforms keywords into fully researched, humanized, and publication-ready long-form articles.
Friday (星期五)
A dedicated AI programming agent designed for automated code generation, security audits, and intelligent refactoring via the opencode CLI.
Abby Watch
A streamlined time management utility for checking current time and tracking countdowns directly from the command line.
Abby Browser Skill
A natural language interface for browser automation that allows users to navigate, interact, and extract data from the web using simple commands.
Abstract Searcher
A tool for automatically populating BibTeX files with academic abstracts via public APIs and intelligent browser automation.
Anemone Browser
A managed headful Chrome browser designed for agents, featuring anti-detection, human-in-the-loop VNC takeover, and multi-session isolation.
Featured*
