PyMuPDF
The Python PDF workhorse: fast extraction, rendering, and manipulation on top of MuPDF.
About
PyMuPDF is the Python binding for MuPDF, the high-performance document engine maintained by Artifex, and it is the library a lot of serious PDF pipelines are built on. Text extraction is the headline, with layout-aware output and a reputation for speed, but it also renders pages to images, merges and splits documents, edits annotations and form fields, redacts, and converts between formats. It is not PDF-only either: the same API handles XPS, EPUB, and HTML. A Tesseract integration lets scripts OCR image-only pages on demand, and the companion pymupdf4llm package turns documents into clean Markdown for LLM ingestion.
The repo holds around 10,900 stars with commits landing the day this was researched, and documentation is extensive. The catch is licensing: it is AGPL-3.0, which means network-deployed commercial use requires a commercial license from Artifex. That is a deliberate business model rather than a flaw, but teams need to price it in before building on it.
What it can do
- Fast text extraction with layout, word coordinates, and reading order
- Render PDF pages to images at any resolution
- Merge, split, rotate, crop, and reorder documents
- Annotate, redact, and edit form fields programmatically
- Handles PDF, XPS, EPUB, CBZ, and HTML through one API
- Tesseract OCR integration for image-only pages
- pymupdf4llm companion for clean Markdown extraction for LLMs
- Encrypted PDFs, digital signatures, and incremental saves
Setup
pip install pymupdf, then import fitz. Commercial deployments should sort out the Artifex license first.
Prerequisites
Python 3.8+; pip install pulls prebuilt wheels. Optional Tesseract for OCR features.
Use cases
Works with
Trust score — 9/10
Scored 9/10 (verified tier, metrics-v1). Evidence: GitHub API confirms 10,914 stars under AGPL-3.0 with a push on 2026-10-16; Artifex maintains both MuPDF and the bindings; third-party comparisons confirm the speed reputation and the multi-format API. Strengths: the fastest serious PDF library in Python, one API for many formats, deep documentation, actively developed. Docked for: (1) AGPL-3.0 means commercial network use needs a paid license; (2) it is a programmer's tool with no GUI or no-code surface; (3) C dependency makes installs heavier than pure-Python alternatives. License: AGPL-3.0, commercial license available.
★ 10,914 GitHub stars
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 9.0/10
- Maintenance · recency, releases, issue load
- 10.0/10
- Openness · license + docs
- 8.0/10
- Manual review · sanity read + red-flag check
- 9.0/10
- · Artifex; Python bindings for the MuPDF engine
- · ~10,914 GitHub stars; push on 2026-10-16
- · Fast layout-aware text extraction; page rendering to images
- · Merge, split, annotate, redact, form fields, encryption
- · One API for PDF, XPS, EPUB, CBZ, HTML
- · Tesseract OCR integration; pymupdf4llm companion for LLM ingestion
- · AGPL-3.0 copyleft: commercial network deployment needs a paid license
- · Programmer-only tool; C dependency vs pure-Python alternatives
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.