An advanced document parsing skill that extracts structured text, complex tables, and LaTeX formulas with Apple Silicon acceleration.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install mineru-pdf
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-pdf 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 PDF Parser?
MinerU PDF Parser is a robust tool designed to transform static PDF documents into rich, AI-ready Markdown. As a key component of the Openclaw Skills ecosystem, it leverages the MinerU MCP server to handle everything from basic text extraction to the identification of complex mathematical notation and tabular data. It is particularly valuable for researchers and developers who need to ingest technical documentation into AI models without losing structural integrity.
This skill is uniquely optimized for the modern developer environment, offering specialized backends that utilize MLX acceleration for users on Apple Silicon hardware. By providing both a native MCP interface for interactive use and a direct Python tool for persistent storage, it ensures that your Openclaw Skills workflow remains flexible and efficient, whether you are performing a quick query or building a massive document database.
MinerU PDF Parser Use Cases
- Digitizing technical whitepapers into LaTeX and Markdown for research repositories.
- Extracting financial data from multi-page PDF tables into structured formats for analysis.
- Pre-processing scanned documents using built-in OCR to prepare them for RAG pipelines.
- Batch processing large document libraries on Mac hardware using MLX-accelerated inference.
- Converting academic textbooks into structured text while preserving mathematical formulas.
How MinerU PDF Parser Works
- The user initiates a request by providing a PDF path and selecting a processing backend like pipeline or vlm-mlx-engine.
- The system invokes the MinerU engine to perform layout analysis and optical character recognition if required.
- Complex elements such as tables and mathematical equations are isolated and converted into Markdown and LaTeX respectively.
- The skill compiles the processed components into a cohesive Markdown document containing metadata and structured content.
- Results are either returned immediately to the AI agent or saved to a local persistent directory for long-term access.
MinerU PDF Parser Setup
To add this capability to your environment, install the MinerU MCP server using the following command:
claude mcp add --transport stdio --scope user mineru -- \
uvx --from mcp-mineru python -m mcp_mineru.server
For users who require file persistence and wish to avoid automatic cleanup of temporary files, use the direct parsing tool provided within the skill directory:
python /path/to/skills/mineru-pdf/parse.py <pdf_path> <output_dir>
MinerU PDF Parser Data Schema & Taxonomy
The skill produces a structured output directory when used in persistent mode, ensuring all assets are easily accessible for your Openclaw Skills projects:
| Path Component | Description |
|---|---|
input_name_parsed.md |
The main Markdown file containing the full extraction. |
/images/ |
A subdirectory containing all figures and graphical elements extracted from the PDF. |
metadata |
Integrated header containing backend settings, page counts, and processing timestamps. |
LaTeX |
Mathematical content is embedded directly within the Markdown using standard LaTeX delimiters. |
MinerU PDF Parser Advanced Features
- Hardware acceleration support for Apple Silicon (M1-M4) via the MLX engine backend.
- Granular page range selection allowing users to process specific chapters or sections.
- Toggleable recognition modules to disable table or formula extraction for faster processing speeds.
- Support for multiple input formats including PDF, JPEG, PNG, and WebP.
- Persistent storage mode via a direct CLI tool to bypass temporary directory auto-deletion.
SKILL.md
Loading
Related Openclaw Skills
buildlog
The buildlog skill captures your AI-assisted coding sessions in real-time to create replayable, shareable recordings.
Nano Banana Zora Deployer
An automated pipeline to generate Nano Banana style character art and deploy it as an NFT on the Zora Network.
Clawhub Skill Creator
A professional framework and toolkit designed to streamline the development, validation, and publication of AI-optimized Openclaw Skills.
rtk-compress
A high-performance CLI proxy skill that wraps commands to filter and compress output, saving up to 90% of LLM tokens.
email-otp
A specialized tool for generating temporary email addresses and automatically capturing verification codes or validation links for automated signups.
leak-buy
A specialized skill for purchasing and downloading digital assets through the leak-cli interface within the Openclaw Skills ecosystem.
Featured*
