# PyMuPDF > PyMuPDF is the fastest PDF processing library for Python. Built on the MuPDF C engine by Artifex: PyMuPDF (general-purpose PDF manipulation and LLM and RAG-optimized extraction with Markdown/JSON output), and PyMuPDF Office (commercial-only enterprise edition with Microsoft Office and Hangul format support). All three share one foundation, one API, and one extraction engine — choose by output format and license model. PyMuPDF covers the full document-processing surface: extraction, analysis, conversion, and manipulation with structure-aware Markdown output, automatic OCR, and native LangChain/LlamaIndex integration — using a Graph Neural Network trained on PDF internals (not pixels) to deliver vision-model accuracy on CPU only. Key facts: - Built on MuPDF, the C engine with 20+ years of PDF heritage. - Includes Layout — GNN-based document intelligence, no GPU required. - AGPL-licensed for open source use; commercial licenses available for proprietary deployment. - Native integrations with LangChain, LlamaIndex. ## Products - [PyMuPDF](https://pymupdf.io/pymupdf): The fastest PDF processing library for Python. Lightning-fast PDF manipulation with minimal dependencies. AGPL or commercial. Use for extraction, conversion, manipulation, and document analysis. `pip install PyMuPDF` - [PyMuPDF Office](https://pymupdf.io/office): Commercial-only enterprise edition. Adds Microsoft Office (DOC/DOCX, XLS/XLSX, PPT/PPTX) and Hangul (HWP/HWPX) input via a conversion layer. Includes everything in PyMuPDF. - [PyMuPDF Office Trial](https://pymupdf.io/office/try) ## Documentation - [PyMuPDF documentation](https://pymupdf.readthedocs.io/en/latest/): Full API reference, recipes, and guides for the core PyMuPDF library. - [Get started](https://pymupdf.io/get-started): Onboarding entry point for new users. ## Recipes and tutorials - [Extract all document text](https://pymupdf.readthedocs.io/en/latest/recipes-text.html#how-to-extract-all-document-text) - [Extract text in natural reading order](https://pymupdf.readthedocs.io/en/latest/recipes-text.html#how-to-extract-text-in-natural-reading-order) - [Extract table content from documents](https://pymupdf.readthedocs.io/en/latest/recipes-text.html#how-to-extract-table-content-from-documents) - [Mark extracted text](https://pymupdf.readthedocs.io/en/latest/recipes-text.html#how-to-mark-extracted-text) ## Distribution - [PyMuPDF on PyPI](https://pypi.org/project/PyMuPDF/): `pip install pymupdf` ## Capabilities - [Extraction](https://pymupdf.io/pymupdf): Text, images, tables, metadata, fonts, annotations. - [Analysis](https://pymupdf.io/pymupdf): Reading order, document structure, layout detection, table identification. - [Conversion](https://pymupdf.io/pymupdf): PDF to Markdown, HTML, images, or text. - [Manipulation](https://pymupdf.io/pymupdf): Merge, split, rotate, crop, watermark, annotate, edit pages. ## Licensing and contact - [Licensing](https://pymupdf.io/licensing): Commercial licensing for AGPL-incompatible use cases. - [Contact](https://artifex.com/contact): Sales and support inquiries. - [Privacy Policy](https://pymupdf.io/privacy-policy) ## Optional - [Blog](https://pymupdf.io/blog): Technical articles, release notes, and tutorials. - [PDF4LLM: The Pre-LLM Document Processing Layer](https://pymupdf.io/blog/pdf4llm-the-pre-llm-document-processing-layer): Positioning piece on PDF4LLM (April 2026). - [Hybrid OCR in PyMuPDF4LLM](https://pymupdf.io/blog/hybrid-ocr-in-pymupdf4llm): Hybrid text/OCR strategy explained (March 2026). - [Introducing the New PyMuPDF4LLM: Now Including Layout](https://pymupdf.io/blog/introducing-the-new-pymupdf4llm-now-including-layout): GNN-based layout release announcement (March 2026). - [Forum](https://forum.mupdf.com/c/general/4): Community forum for PyMuPDF and MuPDF developers. - [PDF4LLM marketing site](https://www.pdf4llm.com): Dedicated landing page for the LLM-extraction product family. - [MuPDF](https://mupdf.com): The C/C++ engine that powers PyMuPDF. - [Artifex](https://artifex.com): Parent company. ## Benchmark - [Benchmark](https://pymupdf.io/benchmark): Losslessly compress your existing PDF archive on the PyMuPDF stack you already run. No migration, no new format, just a smaller bill. ## Compress - [Compress](https://pymupdf.io/compress): Losslessly compress your existing PDF archive on the PyMuPDF stack you already run. No migration, no new format, just a smaller bill. ## Compare Docling - [Compare with Docling](https://pymupdf.io/compare/docling): Compare PyMuPDF and Docling in terms of install weight, speed, ecosystem, and more.