GitHub - opendatalab/MinerU: Transforms complex documents like PDFs and Office docs into LLM-ready markdown/JSON for your Agentic workflows. (original) (raw)

MinerU — High-accuracy document parsing engine for LLM · RAG · Agent workflowsConverts PDF · DOCX · PPTX · XLSX · Images · Web pages into structured Markdown / JSON · VLM+OCR dual engine · 109 languages
MCP Server · LangChain / Dify / FastGPT native integration · 10+ domestic AI chip support

🔍 Core Parsing Capabilities

• Native support for DOCX, PPTX, and XLSX parsing
• Formulas → LaTeX · Tables → HTML, accurate layout reconstruction
• Supports scanned docs, handwriting, multi-column layouts, cross-page table merging
• Output follows human reading order with automatic header/footer removal
• VLM + OCR dual engine, 109-language OCR recognition

🔌 Integration

🖥️ Deployment (Private · Fully Offline)

Domestic AI chips: Ascend · Cambricon · Enflame · MetaX · Moore Threads · Kunlunxin · Iluvatar · Hygon · Biren · T-Head

Changelog

• 2026/06/11 3.3 Released
  This release focuses on Hybrid parsing performance optimization and VLM model capability upgrades. The main updates include:
  • New effort parsing-strength parameter for the Hybrid backend
    * Added two parsing-strength levels, medium and high, allowing users to balance parsing speed, parsing accuracy, and feature requirements.
    * On OmniDocBench v1.6, medium reduces overall accuracy by only 0.13 points compared with high, while delivering 35% ~ 220% parsing speed improvements across different devices and scenarios:
    * Linux: about 80% faster for text PDF scenarios and about 35% faster for OCR scenarios
    * Windows: about 90% faster for text PDF scenarios and about 45% faster for OCR scenarios
    * macOS: about 220% faster for text PDF scenarios and about 50% faster for OCR scenarios
    * The default Hybrid backend now uses effort=medium, significantly improving overall parsing efficiency while maintaining high parsing accuracy.
    * The medium level does not support image analysis; for maximum parsing accuracy or image analysis support, switch to the high-strength parsing mode with effort=high, which may have an impact on parsing speed.
  • VLM model upgraded to MinerU2.5-Pro-2605-1.2B
    * Fixed multiple model issues found in the 2604 version, further improving parsing stability on complex documents.
    * Added native multilingual OCR support, reducing the need for extra language-parameter configuration and improving out-of-the-box usability for multilingual documents.

With the 3.3 release, MinerU further improves Hybrid backend efficiency across platforms and scenarios while maintaining high-accuracy parsing. The default medium effort level is better suited for most day-to-day document processing tasks, while high is designed for scenarios that require maximum parsing accuracy or image analysis capabilities.

• 2026/04/18 3.1.0 Released
  This release focuses on licensing openness, parsing accuracy, and full-format native support. The main updates include:
  • License upgrade
    * MinerU has officially moved from AGPLv3 to the MinerU Open Source License, a custom license based on Apache 2.0.
    * This change significantly reduces adoption friction for both community users and commercial deployments, making MinerU easier to integrate into real-world workflows.
  • VLM main model upgrade
    * The primary VLM model has been upgraded to MinerU2.5-Pro-2604-1.2B, bringing overall parsing accuracy to a state-of-the-art level.
    * The new model now supports image and chart parsing, truncated paragraph merging, cross-page table merging, and image recognition inside tables, further strengthening performance on complex document layouts.
  • Full-format native parsing support
    * Native parsing support has now been extended to PPTX and XLSX.
    * MinerU now fully supports parsing across images, PDF, DOCX, PPTX, and XLSX, providing a more complete multi-format document understanding workflow.

With the 3.1.0 release, MinerU becomes more open, more accurate, and easier to adopt in production. The new license lowers the barrier for both community and commercial use, MinerU2.5-Pro-2604-1.2B improves parsing quality on complex content, and native PPTX / XLSX support completes end-to-end coverage of mainstream document formats.

• 2026/03/29 3.0.0 Released
  This release delivers a systematic upgrade centered on parsing capability, system architecture, and engineering usability. The main updates include:
  • Native DOCX parsing
    * Official support for native DOCX parsing, delivering high-precision results without hallucinations.
    * Compared with the traditional workflow of first converting DOCX to PDF and then parsing it, end-to-end speed is improved by tens of times, making it better suited for scenarios with high requirements for both accuracy and throughput.
  • pipeline backend upgrade
    * The pipeline backend achieves a score of 86.2 on OmniDocBench (v1.5), surpassing the accuracy of the previous-generation mainstream VLM MinerU2.0-2505-0.9B.
    * Added support for parsing images/formulas inside tables, seal text recognition, vertical text support, and interline formula numbering recognition, continuously improving parsing quality for complex document scenarios.
    * While maintaining high accuracy, it keeps resource usage extremely low and continues to support inference in pure CPU environments.
  • API / CLI / Router orchestration upgrade
    * mineru now runs as an orchestration client based on mineru-api; when --api-url is not provided, it will automatically start a local temporary service.
    * mineru-api adds a new asynchronous task endpoint POST /tasks, supporting task submission, status querying, and result retrieval; meanwhile, it retains the synchronous parsing endpoint POST /file_parse for compatibility with legacy plugins.
    * Added mineru-router, designed for unified entry deployment and task routing across multiple services and multiple GPUs; its interfaces are fully compatible with mineru-api and support automatic task load balancing.
  • Deployment and usability improvements
    * Resolved compatibility issues with torch >= 2.8; the base image has been upgraded to vllm0.11.2 + torch2.9.0, unifying installation paths across different Compute Capabilities.
    * Optimized the parsing pipeline with a sliding-window mechanism, significantly reducing peak memory usage in long-document scenarios, so documents with tens of thousands of pages no longer need to be split manually.
    * Batch inference in pipeline now supports streaming writes to disk, allowing completed parsing results to be written out in time and further improving the experience for long-running tasks.
    * Completed thread-safety optimization and now fully supports multi-threaded concurrent inference; together with mineru-router, this enables one-click multi-GPU deployment and makes it easy to build high-concurrency, high-throughput parsing systems.
    * Completely removed the use of two AGPLv3 models (doclayoutyolo and mfd_yolov8) and one CC-BY-NC-SA 4.0 model (layoutreader).

This update is not just a set of feature enhancements, but a key leap forward in MinerU's overall system capabilities. We specifically addressed the peak memory usage issue in long-document parsing. Through optimizations such as sliding windows and streaming writes to disk, ultra-long document parsing has moved from “requiring manual splitting and careful handling” to being “stable, scalable, and ready for production workloads.” At the same time, we completed thread-safety optimization and fully enabled multi-threaded concurrent inference, further improving single-machine resource utilization and runtime stability under high-concurrency workloads. On top of this, with mineru-router and the new API / CLI orchestration framework, MinerU now supports one-click multi-GPU deployment, unified access across multiple services, and automatic task load balancing, significantly reducing the difficulty of large-scale deployment. As a result, MinerU is evolving from a standalone data production tool into a large-scale document parsing foundation for high-concurrency and high-throughput scenarios, providing enterprise-grade document data processing with infrastructure that is more stable, more efficient, and easier to scale.

📝 View the complete Changelog for more historical version information

Project Introduction

MinerU is a document parsing tool that converts PDF, image, DOCX, PPTX, and XLSX inputs into machine-readable formats such as Markdown and JSON for downstream retrieval, extraction, and processing. MinerU was born during the pre-training process of InternLM. We focus on solving symbol conversion issues in scientific literature and hope to contribute to technological development in the era of large models. Compared to well-known commercial products, MinerU is still young. If you encounter any issues or if the results are not as expected, please submit an issue on issue and attach the relevant document or sample file.

pdf_zh_cn.mp4

Key Features

• Support PDF, image, DOCX, PPTX, and XLSX inputs.
• Remove headers, footers, footnotes, page numbers, etc., to ensure semantic coherence.
• Output text in human-readable order, suitable for single-column, multi-column, and complex layouts.
• Preserve the structure of the original document, including headings, paragraphs, lists, etc.
• Extract images, image descriptions, tables, table titles, and footnotes.
• Automatically recognize and convert formulas in the document to LaTeX format.
• Automatically recognize and convert tables in the document to HTML format.
• Automatically detect scanned PDFs and garbled PDFs and enable OCR functionality.
• OCR supports detection and recognition of 109 languages.
• Supports multiple output formats, such as multimodal and NLP Markdown, JSON sorted by reading order, and rich intermediate formats.
• Supports various visualization results, including layout visualization and span visualization, for efficient confirmation of output quality.
• Built-in CLI, FastAPI, Gradio WebUI, for local orchestration and multi-service deployment.
• Supports running in a pure CPU environment, and also supports GPU/MPS acceleration
• Compatible with Windows, Linux, and Mac platforms.

Quick Start

Document parsing is a difficult and complex task. In scenarios such as complex layouts, scanned pages, and handwritten content, the parsing results may fall short of expectations. We recommend trying the online demo first to evaluate MinerU's parsing quality and suitability before choosing an appropriate deployment method based on your actual needs. If you have document samples with unsatisfactory parsing results, feel free to share them in an issue. We will continue improving the parsing capabilities. If you encounter any installation issues, please first consult the FAQ.

Online Experience

Official online web application

The official online version has the same functionality as the client, with a beautiful interface and rich features, requires login to use

Gradio-based online demo

A WebUI developed based on Gradio, with a simple interface and only core parsing functionality, no login required

Local Deployment

Warning

Pre-installation Notice—Hardware and Software Environment Support

To ensure the stability and reliability of the project, we only optimize and test for specific hardware and software environments during development. This ensures that users deploying and running the project on recommended system configurations will get the best performance with the fewest compatibility issues.

By focusing resources on the mainline environment, our team can more efficiently resolve potential bugs and develop new features.

In non-mainline environments, due to the diversity of hardware and software configurations, as well as third-party dependency compatibility issues, we cannot guarantee 100% project availability. Therefore, for users who wish to use this project in non-recommended environments, we suggest carefully reading the documentation and FAQ first. Most issues already have corresponding solutions in the FAQ. We also encourage community feedback to help us gradually expand support.

1 Accuracy metrics are the End-to-End Evaluation Overall scores from OmniDocBench (v1.6), based on the latest version of MinerU.
2 Servers compatible with OpenAI API, such as local model servers or remote model services deployed via inference frameworks like vLLM/SGLang/LMDeploy.
3 Linux only supports distributions from 2019 and later.
4 Since the key dependency ray does not support Python 3.13 on Windows, only versions 3.10~3.12 are supported.
5 macOS requires version 14.0 or later.

Install MinerU

Install MinerU using pip or uv

pip install --upgrade pip pip install uv uv pip install -U "mineru[all]"

Install MinerU from source code

git clone https://github.com/opendatalab/MinerU.git cd MinerU uv pip install -e .[all]

Tip

• mineru[all] includes all core features, compatible with Windows / Linux / macOS systems, suitable for most users.
• If CUDA acceleration is unavailable after installing on Windows, see the Windows CUDA acceleration FAQ.
• If you need to specify the inference framework for the VLM model, or only intend to install a lightweight client on an edge device, please refer to the documentation Extension Modules Installation Guide.

Deploy MinerU using Docker

MinerU provides a convenient Docker deployment method, which helps quickly set up the environment and solve some tricky environment compatibility issues.

Tip

• Docker deployment is only supported on Linux and Windows environments with WSL2 support;
• macOS users should refer to the two installation methods above for installation instead of using Docker deployment.

You can get the Docker Deployment Instructions in the documentation.

Using MinerU

If your device meets the GPU acceleration requirements in the table above, you can use a simple command line for document parsing:

mineru -p -o

If your device does not meet the GPU acceleration requirements, you can specify the backend as pipeline to run in a pure CPU environment:

mineru -p -o -b pipeline

mineru currently supports local PDF, image, DOCX, PPTX, and XLSX file or directory inputs, and can be used for document parsing through the CLI, API, WebUI, and mineru-router. For detailed instructions, please refer to the Usage Guide.

FAQ

• If you encounter any issues during usage, you can first check the FAQ for solutions.
• If your issue remains unresolved, you may also use DeepWiki to interact with an AI assistant, which can address most common problems.
• If you still cannot resolve the issue, you are welcome to join our community via Discord or WeChat to discuss with other users and developers.

All Thanks To Our Contributors

License Information

This repository is licensed under the MinerU Open Source License, based on Apache 2.0 with additional conditions.

Acknowledgments

• UniMERNet
• TableStructureRec
• PaddleOCR
• PaddleOCR2Pytorch
• fast-langdetect
• pypdfium2
• pdftext
• pypdf
• magika
• vLLM
• LMDeploy

Citation

@article{wang2026mineru2, title={MinerU2. 5-Pro: Pushing the Limits of Data-Centric Document Parsing at Scale}, author={Wang, Bin and He, Tianyao and Ouyang, Linke and Wu, Fan and Zhao, Zhiyuan and Chu, Tao and Qu, Yuan and Jin, Zhenjiang and Zeng, Weijun and Miao, Ziyang and others}, journal={arXiv preprint arXiv:2604.04771}, year={2026} }

@article{dong2026minerudiffusion, title={MinerU-Diffusion: Rethinking Document OCR as Inverse Rendering via Diffusion Decoding}, author={Dong, Hejun and Niu, Junbo and Wang, Bin and Zeng, Weijun and Zhang, Wentao and He, Conghui}, journal={arXiv preprint arXiv:2603.22458}, year={2026} }

@article{niu2025mineru2, title={Mineru2. 5: A decoupled vision-language model for efficient high-resolution document parsing}, author={Niu, Junbo and Liu, Zheng and Gu, Zhuangcheng and Wang, Bin and Ouyang, Linke and Zhao, Zhiyuan and Chu, Tao and He, Tianyao and Wu, Fan and Zhang, Qintong and others}, journal={arXiv preprint arXiv:2509.22186}, year={2025} }

@article{wang2024mineru, title={Mineru: An open-source solution for precise document content extraction}, author={Wang, Bin and Xu, Chao and Zhao, Xiaomeng and Ouyang, Linke and Wu, Fan and Zhao, Zhiyuan and Xu, Rui and Liu, Kaiwen and Qu, Yuan and Shang, Fukai and others}, journal={arXiv preprint arXiv:2409.18839}, year={2024} }

@article{he2024opendatalab, title={Opendatalab: Empowering general artificial intelligence with open datasets}, author={He, Conghui and Li, Wei and Jin, Zhenjiang and Xu, Chao and Wang, Bin and Lin, Dahua}, journal={arXiv preprint arXiv:2407.13773}, year={2024} }

Star History

Links

• MinerU-Diffusion: Rethinking Document OCR as Inverse Rendering via Diffusion Decoding
• Easy Data Preparation with latest LLMs-based Operators and Pipelines
• Vis3 (OSS browser based on s3)
• LabelU (A Lightweight Multi-modal Data Annotation Tool)
• LabelLLM (An Open-source LLM Dialogue Annotation Platform)
• PDF-Extract-Kit (A Comprehensive Toolkit for High-Quality PDF Content Extraction)
• OmniDocBench (A Comprehensive Benchmark for Document Parsing and Evaluation)
• Magic-HTML (Mixed web page extraction tool)
• Magic-Doc (Fast speed ppt/pptx/doc/docx/pdf extraction tool)
• Dingo: A Comprehensive AI Data Quality Evaluation Tool