School

MinerU: Convert PDFs to Markdown for AI

Translated

Before you upload textbooks or research papers to an AI chatbot, convert your PDF and Word documents into Markdown format so the AI can read them more easily.

3-Line Summary

1) It converts PDF and Word documents into Markdown (a text formatting language using symbols) that’s easy for AI to read. 2) You can try it out first on the HuggingFace web demo without installing anything. 3) Use it to organize textbooks and research papers before feeding them into an AI chatbot, and convert Word files in the same way.

Image source: GitHub repository card (opendatalab/MinerU)

At a Glance

• 83,076 stars (as of November 4, 2026) • Programming languages used: Python 99.0% (an easy-to-read, general-purpose language), JavaScript 0.8% (a language for building web interfaces) • License: MinerU Open Source License (a license based on the Apache 2.0 license with additional terms; large companies with over 100 million monthly users require a separate agreement) • Korean: The README is available only in English and Chinese; support for Korean documentation cannot be confirmed https://github.com/opendatalab/MinerU

How to Use It

1) The easiest way is to use the HuggingFace web demo described in the README. Due to server limitations, you can only use the "flash" and "basic" levels. 2) Your computer needs Python 3.11 or later and uv (a Python package installer). Create a virtual environment, then run the following command: uv pip install -U "mineru>=4.0,<5" This command downloads and installs MinerU. 3) Use the command below to convert a PDF to a Markdown file. mineru-kit parse document.pdf -o document.md --tier standard This command reads the PDF and saves it as a Markdown file using the “standard” tier. The “standard” tier requires a graphics card with at least 8 GB of VRAM; if you don’t have one, use the “basic” tier to run it on the CPU.

Popular Comments

GitHub Discussion #4682 · 1 reaction · vectornguyen76 · November 2026 "It's already supported. You just need to upgrade; there's no need to convert to PDF first. Starting with version 4.0, it supports DOC/DOCX, PPT/PPTX, XLS/XLSX, and more." (Translation) https://github.com/opendatalab/MinerU/discussions/4682#discussioncomment-19352053

Points to Note

By default, documents are not uploaded externally, but if you manually enable remote processing, they will be uploaded, so please be careful with personal information and copyrighted material. Please check the conversion results by comparing them with the original text.

This post was organized with the help of AI.

People who tried this

Nobody has tried it yet

Even one result photo is enough. Be the first to share.