opendatalab/MinerU

What is it?

Key features

• 01Parses PDFs, scanned images, Office files, EPUB, HTML and CSV into Markdown or JSON.
• 02Offers four quality tiers from fast Flash parsing to Advanced parsing for complex layouts.
• 03Exports to nine targets including Markdown, HTML, LaTeX, DOCX and EPUB.
• 04Runs local models on CPU, GPU or Apple Silicon, with remote parsing only when configured.
• 05Provides a CLI, Python SDK, API and Gradio WebUI, plus a local document library with search.

Use cases

• Feeding PDF reports or academic papers into an LLM or RAG pipeline as clean Markdown.
• Extracting tables and formulas from scanned documents for analysis.
• Letting AI agents read long documents page by page with citable locators.
• Batch converting Office and OpenDocument files into structured text.

Quick start

1. 1

   Install uv with pip

   pip install uv
2. 2

   Create a virtual environment with Python 3.12

   uv venv .mineru --python 3.12
3. 3

   Install MinerU 4.0 stable into the environment (activate the venv first, per the README)

   uv pip install -U "mineru>=4.0,<5"
4. 4

   Parse a PDF to Markdown

   mineru-kit parse document.pdf -o document.md --tier standard

Things to know

• !The license could not be identified automatically, so check the repository before commercial use.
• !Best throughput needs an NVIDIA GPU; other non-NVIDIA devices need a manually installed accelerated torch and vllm or lmdeploy.
• !On Windows, the GPU build of torch must be installed separately.

Topics

FAQ

What is MinerU?

Open-source document parser that converts PDFs, Office files and images into Markdown or JSON for LLM and agent workflows. MinerU parses complex documents such as PDFs, scanned images, Word, PowerPoint, Excel, EPUB, HTML and CSV into structured content. Output can be rendered as Markdown, HTML, LaTeX, DOCX, EPUB, PDF, or structured JSON content lists, which makes it suitable as input for RAG pipelines and LLM applications.

Is MinerU free to use? What license does it have?

MinerU is open source under the Other license. Check the license terms before using it in commercial products.

Is MinerU still maintained?

Actively maintained. The last commit to MinerU was on Oct 19, 2026.

What can I use MinerU for?

Feeding PDF reports or academic papers into an LLM or RAG pipeline as clean Markdown. Extracting tables and formulas from scanned documents for analysis. Letting AI agents read long documents page by page with citable locators. Batch converting Office and OpenDocument files into structured text.

How do I install MinerU?

Install uv with pip: pip install uv Create a virtual environment with Python 3.12: uv venv .mineru --python 3.12 Install MinerU 4.0 stable into the environment (activate the venv first, per the README): uv pip install -U "mineru>=4.0,<5" Parse a PDF to Markdown: mineru-kit parse document.pdf -o document.md --tier standard

Where can I get MinerU?

The source code and installation instructions are on GitHub: https://github.com/opendatalab/MinerU