• 2025/05/24 1.3.12 Released
  • Added support for ppocrv5 model, updated ch_server model to PP-OCRv5_rec_server and ch_lite model to PP-OCRv5_rec_mobile (model update required)
    • In testing, we found that ppocrv5(server) shows some improvement for handwritten documents, but slightly lower accuracy than v4_server_doc for other document types. Therefore, the default ch model remains unchanged as PP-OCRv4_server_rec_doc.
    • Since ppocrv5 enhances recognition capabilities for handwritten text and special characters, you can manually select ppocrv5 models for Japanese, traditional Chinese mixed scenarios and handwritten document scenarios
    • You can select the appropriate model through the lang parameter lang='ch_server' (python api) or --lang ch_server (command line):
      • ch: PP-OCRv4_rec_server_doc (default) (Chinese, English, Japanese, Traditional Chinese mixed/15k dictionary)
      • ch_server: PP-OCRv5_rec_server (Chinese, English, Japanese, Traditional Chinese mixed + handwriting/18k dictionary)
      • ch_lite: PP-OCRv5_rec_mobile (Chinese, English, Japanese, Traditional Chinese mixed + handwriting/18k dictionary)
      • ch_server_v4: PP-OCRv4_rec_server (Chinese, English mixed/6k dictionary)
      • ch_lite_v4: PP-OCRv4_rec_mobile (Chinese, English mixed/6k dictionary)
  • Added support for handwritten documents by optimizing layout recognition of handwritten text areas
    • This feature is supported by default, no additional configuration needed
    • You can refer to the instructions above to manually select ppocrv5 model for better handwritten document parsing
  • The demos on huggingface and modelscope have been updated to support handwriting recognition and ppocrv5 models, which you can experience online
• 2025/04/29 1.3.10 Released
  • Support for custom formula delimiters can be achieved by modifying the latex-delimiter-config item in the magic-pdf.json file under the user directory.
• 2025/04/27 1.3.9 Released
  • Optimized the formula parsing function to improve the success rate of formula rendering
• 2025/04/23 1.3.8 Released
  • The default ocr model (ch) has been updated to PP-OCRv4_server_rec_doc (model update required)
    • PP-OCRv4_server_rec_doc is trained on a mix of more Chinese document data and PP-OCR training data, enhancing recognition capabilities for some traditional Chinese characters, Japanese, and special characters. It supports over 15,000 recognizable characters, improving text recognition in documents while also boosting general text recognition.
    • Performance comparison between PP-OCRv4_server_rec_doc, PP-OCRv4_server_rec, and PP-OCRv4_mobile_rec
    • Verified results show that the PP-OCRv4_server_rec_doc model significantly improves accuracy in both single-language (Chinese, English, Japanese, Traditional Chinese) and mixed-language scenarios, with speed comparable to PP-OCRv4_server_rec, making it suitable for most use cases.
    • In a small number of pure English scenarios, the PP-OCRv4_server_rec_doc model may encounter word concatenation issues, whereas PP-OCRv4_server_rec performs better in such cases. Therefore, we have retained the PP-OCRv4_server_rec model, which users can invoke by passing the parameter lang='ch_server'(python api) or --lang ch_server(cli).
• 2025/04/22 1.3.7 Released
  • Fixed the issue where the lang parameter was ineffective during table parsing model initialization.
  • Fixed the significant slowdown in OCR and table parsing speed in cpu mode.
• 2025/04/16 1.3.4 Released
  • Slightly improved the speed of OCR detection by removing some unused blocks.
  • Fixed page-level sorting errors caused by footnotes in certain cases.
• 2025/04/12 1.3.2 released
  • Fixed the issue of incompatible dependency package versions when installing in Python 3.13 environment on Windows systems.
  • Optimized memory usage during batch inference.
  • Improved the parsing effect of tables rotated by 90 degrees.
  • Enhanced the parsing accuracy for large tables in financial report samples.
  • Fixed the occasional word concatenation issue in English text areas when OCR language is not specified.(The model needs to be updated)
• 2025/04/08 1.3.1 released, fixed some compatibility issues
  • Supported Python 3.13
  • Made the final adaptation for some outdated Linux systems (e.g., CentOS 7), and no further support will be guaranteed for subsequent versions. Installation Instructions
• 2025/04/03 1.3.0 released, in this version we made many optimizations and improvements:
  • Installation and compatibility optimization
    • By removing the use of layoutlmv3 in layout, resolved compatibility issues caused by detectron2.
    • Torch version compatibility extended to 2.2~2.6 (excluding 2.5).
    • CUDA compatibility supports 11.8/12.4/12.6/12.8 (CUDA version determined by torch), resolving compatibility issues for some users with 50-series and H-series GPUs.
    • Python compatible versions expanded to 3.10~3.12, solving the problem of automatic downgrade to 0.6.1 during installation in non-3.10 environments.
    • Offline deployment process optimized; no internet connection required after successful deployment to download any model files.
  • Performance optimization
    • By supporting batch processing of multiple PDF files (script example), improved parsing speed for small files in batches (compared to version 1.0.1, formula parsing speed increased by over 1400%, overall parsing speed increased by over 500%).
    • Optimized loading and usage of the mfr model, reducing GPU memory usage and improving parsing speed (requires re-execution of the model download process to obtain incremental updates of model files).
    • Optimized GPU memory usage, requiring only a minimum of 6GB to run this project.
    • Improved running speed on MPS devices.
  • Parsing effect optimization
    • Updated the mfr model to unimernet(2503), solving the issue of lost line breaks in multi-line formulas.
  • Usability Optimization
    • By using paddleocr2torch, completely replaced the use of the paddle framework and paddleocr in the project, resolving conflicts between paddle and torch, as well as thread safety issues caused by the paddle framework.
    • Added a real-time progress bar during the parsing process to accurately track progress, making the wait less painful.

MinerU is a tool that converts PDFs into machine-readable formats (e.g., markdown, JSON), allowing for easy extraction into any format. 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 PDF.

https://github.com/user-attachments/assets/4bea02c9-6d54-4cd6-97ed-dff14340982c

• 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 84 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.
• Supports running in a pure CPU environment, and also supports GPU(CUDA)/NPU(CANN)/MPS acceleration
• Compatible with Windows, Linux, and Mac platforms.

If you encounter any installation issues, please first consult the FAQ.
If the parsing results are not as expected, refer to the Known Issues.
There are three different ways to experience MinerU:

• Online Demo (No Installation Required)
• Quick CPU Demo (Windows, Linux, Mac)
• Accelerate inference by using CUDA/CANN/MPS
  • Linux/Windows + CUDA
  • Linux + CANN
  • MacOS + MPS

Synced with dev branch updates:

conda create -n mineru 'python=3.12' -y
conda activate mineru
pip install -U "magic-pdf[full]"

Refer to How to Download Model Files for detailed instructions.

After completing the 2. Download model weight files step, the script will automatically generate a magic-pdf.json file in the user directory and configure the default model path. You can find the magic-pdf.json file in your 【user directory】.

You can modify certain configurations in this file to enable or disable features, such as table recognition:

{
    // other config
    "layout-config": {
        "model": "doclayout_yolo" 
    },
    "formula-config": {
        "mfd_model": "yolo_v8_mfd",
        "mfr_model": "unimernet_small",
        "enable": true  // The formula recognition feature is enabled by default. If you need to disable it, please change the value here to "false".
    },
    "table-config": {
        "model": "rapid_table", 
        "sub_model": "slanet_plus",
        "enable": true, // The table recognition feature is enabled by default. If you need to disable it, please change the value here to "false".
        "max_time": 400
    }
}

If your device supports CUDA and meets the GPU requirements of the mainline environment, you can use GPU acceleration. Please select the appropriate guide based on your system:

• Ubuntu 22.04 LTS + GPU
• Windows 10/11 + GPU
• Quick Deployment with Docker

docker run --rm --gpus=all nvidia/cuda:12.1.0-base-ubuntu22.04 nvidia-smi

wget https://github.com/opendatalab/MinerU/raw/master/docker/global/Dockerfile -O Dockerfile
docker build -t mineru:latest .
docker run -it --name mineru --gpus=all mineru:latest /bin/bash -c "echo 'source /opt/mineru_venv/bin/activate' >> ~/.bashrc && exec bash"
magic-pdf --help

If your device has NPU acceleration hardware, you can follow the tutorial below to use NPU acceleration:

Ascend NPU Acceleration

If your device uses Apple silicon chips, you can enable MPS acceleration for your tasks.

You can enable MPS acceleration by setting the device-mode parameter to mps in the magic-pdf.json configuration file.

{
    // other config
    "device-mode": "mps"
}

Using MinerU via Command Line

Using MinerU via Python API

Derived projects include secondary development projects based on MinerU by project developers and community developers,
such as application interfaces based on Gradio, RAG based on llama, web demos similar to the official website, lightweight multi-GPU load balancing client/server ends, etc. These projects may offer more features and a better user experience.
For specific deployment methods, please refer to the Derived Project README

TODO

• Reading order based on the model
• Recognition of index and list in the main text
• Table recognition
• Heading Classification
• Code block recognition in the main text
• Chemical formula recognition
• Geometric shape recognition

• Reading order is determined by the model based on the spatial distribution of readable content, and may be out of order in some areas under extremely complex layouts.
• Vertical text is not supported.
• Tables of contents and lists are recognized through rules, and some uncommon list formats may not be recognized.
• Code blocks are not yet supported in the layout model.
• Comic books, art albums, primary school textbooks, and exercises cannot be parsed well.
• Table recognition may result in row/column recognition errors in complex tables.
• OCR recognition may produce inaccurate characters in PDFs of lesser-known languages (e.g., diacritical marks in Latin script, easily confused characters in Arabic script).
• Some formulas may not render correctly in Markdown.

FAQ in Chinese

FAQ in English

LICENSE.md

This project currently uses PyMuPDF to achieve advanced functionality. However, since it adheres to the AGPL license, it may impose restrictions on certain usage scenarios. In future iterations, we plan to explore and replace it with a more permissive PDF processing library to enhance user-friendliness and flexibility.

• PDF-Extract-Kit
• DocLayout-YOLO
• StructEqTable
• RapidTable
• PaddleOCR
• RapidOCR
• PaddleOCR2Pytorch
• PyMuPDF
• layoutreader
• fast-langdetect
• pdfminer.six

@misc{wang2024mineruopensourcesolutionprecise,
      title={MinerU: An Open-Source Solution for Precise Document Content Extraction}, 
      author={Bin Wang and Chao Xu and Xiaomeng Zhao and Linke Ouyang and Fan Wu and Zhiyuan Zhao and Rui Xu and Kaiwen Liu and Yuan Qu and Fukai Shang and Bo Zhang and Liqun Wei and Zhihao Sui and Wei Li and Botian Shi and Yu Qiao and Dahua Lin and Conghui He},
      year={2024},
      eprint={2409.18839},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2409.18839}, 
}

@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}
}

Magic-Doc Fast speed ppt/pptx/doc/docx/pdf extraction tool

Magic-HTML Mixed web page extraction tool

• 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)