MinerU review

Q: Is MinerU free and open source?

MinerU's code is on GitHub (opendatalab/MinerU), but the repository doesn't declare a standard open-source license. Check its license file before using it commercially.

Q: How do I install MinerU?

Run uv pip install -U "mineru>=4.0,<5". The mineru package needs Python 3.10 or newer, below 3.15. The install section above has the other options from the project's README.

Q: Is MinerU still maintained?

Yes. The opendatalab/MinerU repository was last updated on Oct 25, 2026, and the latest release, mineru-4.0.12-released on GitHub, came out on Oct 11, 2026.

Q: How popular is MinerU?

MinerU has 82.9k GitHub stars as of Nov 1, 2026. It gained 517 new stars in the week of Oct 25–31, 2026, against a 12-week average of 522 a week. Its main package had 30.8k downloads that week, +20% against its own 4-week average. That made it #11 among RAG, data and memory tools on the Devtoolwatch Board.

Q: MinerU vs Docling: how do they compare?

MinerU and Docling are both tracked by Devtoolwatch as RAG, data and memory tools. MinerU has 82.9k GitHub stars and Docling has 69.9k. In the week of Oct 25–31, 2026, MinerU's main package had 30.8k downloads and Docling's had 904.2k. Docling is a Python SDK and CLI that parses PDF, DOCX, HTML and other formats into a unified document representation for gen AI.

Q: MinerU vs Marker: how do they compare?

MinerU and Marker are both tracked by Devtoolwatch as RAG, data and memory tools. MinerU has 82.9k GitHub stars and Marker has 40.7k. In the week of Oct 25–31, 2026, MinerU's main package had 30.8k downloads and Marker's had 20.6k.

Q: What are the best alternatives to MinerU?

Other RAG, data and memory tools on the Devtoolwatch Board in the week of Oct 25–31, 2026, by momentum: XERJ (#1), Memvid (#2), LEANN (#3), Graphify (#4), MarkItDown (#5). The alternatives table above has their numbers.

Q: How do I use MinerU in Python?

MinerU provides DoclibClient for Python. Start the server with mineru server start, call ensure_parse with a ParseRequest, poll get_parse until done, then use read_content with a locator such as doc:{short_id}/tier:standard/page:1.

Q: Does MinerU upload my documents to a server?

MinerU does not automatically upload documents to the official service. Remote parsing requires explicit configuration, and the agent guide says to ask before using --remote.

Q: Can MinerU handle long documents?

MinerU supports page and block continuation. mineru parse --json defaults to the first 10 PDF pages and returns locators and next_request for continued reads. Stateless mineru-kit parse defaults to all pages.

MinerU is a Python document parser that converts PDFs, images, and Office files into Markdown or JSON for LLM and agent workflows.

Updated Nov 1, 2026. Review written Oct 25, 2026 from the project's own sources; usage data refreshed every Monday.

TL;DR

MinerU parses PDFs, scanned images, Office, EPUB, OFD, HTML, and CSV files into Markdown, JSON, and other formats. MinerU is for developers and AI agents that need to read long documents with tables, formulas, or complex layouts. MinerU offers four parsing tiers, a local document library with search, and stable page and block locators for citations.

82.9kGitHub stars
+517new stars, week of Oct 25
30.8kdownloads a week, +20% vs 4-week avg
#11in RAG, data and memory this week

MinerU at a glance

Category: RAG, data and memory
Maintained by: OpenDataLab (organization, @opendatalab on GitHub)
Language: Python, JavaScript, Shell
Repository: github.com/opendatalab/MinerU
Website: opendatalab.github.io/MinerU
PyPI package: mineru 4.0.12
Latest release: mineru-4.0.12-released on GitHub, Oct 11, 2026
First published: Feb 29, 2024

What MinerU does

MinerU parses PDF, images, DOC/DOCX, PPT/PPTX, XLS/XLSX, RTF, OpenDocument, EPUB, OFD, HTML/MHTML, and CSV/TSV files.
MinerU offers four parsing tiers: Flash for fast previews, Basic for OCR, and Standard and Advanced for demanding layouts.
MinerU renders one document model to targets including Markdown, HTML, LaTeX, DOCX, EPUB, PDF, and Content List V1/V2.
MinerU keeps a local document library that discovers files, caches results, and searches indexed content.
MinerU returns stable doc/tier/page/block locators so agents can continue reading and cite exact locations.
MinerU can retrieve page or block images for visual inspection.

How it works

MinerU runs locally through the mineru CLI, stateless mineru-kit batch conversion, a Python SDK, a V1 HTTP API, a multi-service Router, and a Gradio WebUI. Small models use ONNX or Torch. The VLM uses llama.cpp, vLLM, or LMDeploy. The default install runs ONNX on CPU and llama.cpp in Vulkan mode. Documents are not uploaded to the official service unless remote parsing is configured.

How to install and start using MinerU

With uv

uv pip install -U "mineru>=4.0,<5"

As an agent skill

npx skills add opendatalab/MinerU --skill mineru --global --yes

Requirements

Python >=3.10,<3.15
NVIDIA GPU users should install mineru[full]>=4.0 for best throughput; on Windows the GPU torch build is installed separately.
The Python SDK requires starting the local server first with mineru server start.

First steps

Create a virtual environment with uv and Python 3.12, then install mineru>=4.0,<5.
Convert a file with mineru-kit parse document.pdf -o document.md --tier standard.
Open the browser interface with mineru-kit webui.
For agent reading, run mineru parse "document.pdf" --json and continue with the returned locators.

Who MinerU is for

AI agents that read, quote, and cite local documents with stable page and block references.
Developers converting PDFs with tables, formulas, figures, or complex layouts into Markdown or JSON.
Teams that need OCR on scanned PDFs and images while keeping documents local.

Limitations to know

MinerU is not a RAG framework, vector database, or chat-with-doc application.
Office, OpenDocument, EPUB, OFD, HTML/MHTML, and CSV/TSV files use only the Flash tier; only PDF and images support all four tiers.
Non-NVIDIA accelerated devices need a manually installed accelerated torch plus vllm or lmdeploy for best speed.
Docker deployment for non-NVIDIA devices is pending an update.

Is MinerU growing? Usage data

In the week of Oct 25–31, 2026, MinerU ranked #124 on the Devtoolwatch Board and #11 in RAG, data and memory, with a momentum score of 57 out of 100. Its main package had 30.8k downloads that week, +20% against its own 4-week average.

New GitHub stars per week for the last 12 weeks, from GitHub's stargazer history (today's stargazers by the week they starred).

Show the weekly numbers

New GitHub stars per week for MinerU
Week of	New stars
Aug 9, 2026	596
Aug 16, 2026	564
Aug 23, 2026	545
Aug 30, 2026	571
Sep 6, 2026	536
Sep 13, 2026	541
Sep 20, 2026	455
Sep 27, 2026	423
Oct 4, 2026	412
Oct 11, 2026	562
Oct 18, 2026	542
Oct 25, 2026	517

On Hacker News

Open-source tool helps you convert PDF documents, web pages, etc., into Markdown, 61 points and 3 comments (Aug 17, 2024)

The Board ranks tools by four signals each week: new stars, new stars for the repo's size, added downloads and download growth beyond the market. How the ranking works.

MinerU alternatives

Other RAG, data and memory tools on the Board in the week of Oct 25–31, 2026, by momentum:

RAG, data and memory ranked by momentum, the week of Oct 25–31, 2026
Rank	Tool	Stars	Downloads/wk
#1	XERJXERJ is a local-first search engine and MCP server for AI agents, packaged as one Rust binary with an Elasticsearch-compatible API.	17.9k	–
#2	MemvidMemvid is a single-file memory layer for AI agents that packages data, embeddings, search indexes, and metadata into one portable .mv2 file.	17.3k	–
#3	LEANNLEANN is a Python vector database and RAG library that recomputes embeddings on demand to keep indexes small on personal devices.	13.7k	–
#4	GraphifyGraphify is an AI coding assistant skill that turns a codebase, with its docs, PDFs, images and videos, into a queryable knowledge graph.	140k	480.7k
#5	MarkItDownMarkItDown is a lightweight Python utility from Microsoft that converts PDFs, Office documents, images, audio and other files to…	193.2k	4.7M
#6	FirecrawlFirecrawl is an API and open source project to search, scrape, crawl, and interact with web pages, returning LLM-ready markdown or JSON…	197.4k	2.9M

All RAG, data and memory tools, ranked

Recent MinerU releases

mineru-4.0.12-released, Oct 11, 2026
mineru-4.0.11-released, Oct 8, 2026
mineru-4.0.10-released, Sep 29, 2026
mineru-4.0.9-released, Sep 28, 2026
mineru-4.0.8-released, Sep 27, 2026

MinerU FAQ

Is MinerU free and open source?

MinerU's code is on GitHub (opendatalab/MinerU), but the repository doesn't declare a standard open-source license. Check its license file before using it commercially.

How do I install MinerU?

Run uv pip install -U "mineru>=4.0,<5". The mineru package needs Python 3.10 or newer, below 3.15. The install section above has the other options from the project's README.

Is MinerU still maintained?

Yes. The opendatalab/MinerU repository was last updated on Oct 25, 2026, and the latest release, mineru-4.0.12-released on GitHub, came out on Oct 11, 2026.

How popular is MinerU?

MinerU has 82.9k GitHub stars as of Nov 1, 2026. It gained 517 new stars in the week of Oct 25–31, 2026, against a 12-week average of 522 a week. Its main package had 30.8k downloads that week, +20% against its own 4-week average. That made it #11 among RAG, data and memory tools on the Devtoolwatch Board.

MinerU vs Docling: how do they compare?

MinerU and Docling are both tracked by Devtoolwatch as RAG, data and memory tools. MinerU has 82.9k GitHub stars and Docling has 69.9k. In the week of Oct 25–31, 2026, MinerU's main package had 30.8k downloads and Docling's had 904.2k. Docling is a Python SDK and CLI that parses PDF, DOCX, HTML and other formats into a unified document representation for gen AI. Docling review.

MinerU vs Marker: how do they compare?

MinerU and Marker are both tracked by Devtoolwatch as RAG, data and memory tools. MinerU has 82.9k GitHub stars and Marker has 40.7k. In the week of Oct 25–31, 2026, MinerU's main package had 30.8k downloads and Marker's had 20.6k. Marker review.

What are the best alternatives to MinerU?

Other RAG, data and memory tools on the Devtoolwatch Board in the week of Oct 25–31, 2026, by momentum: XERJ (#1), Memvid (#2), LEANN (#3), Graphify (#4), MarkItDown (#5). The alternatives table above has their numbers.

How do I use MinerU in Python?

MinerU provides DoclibClient for Python. Start the server with mineru server start, call ensure_parse with a ParseRequest, poll get_parse until done, then use read_content with a locator such as doc:{short_id}/tier:standard/page:1.

Does MinerU upload my documents to a server?

MinerU does not automatically upload documents to the official service. Remote parsing requires explicit configuration, and the agent guide says to ask before using --remote.

Can MinerU handle long documents?

MinerU supports page and block continuation. mineru parse --json defaults to the first 10 PDF pages and returns locators and next_request for continued reads. Stateless mineru-kit parse defaults to all pages.

Does MinerU need a GPU?

MinerU works without a GPU by default. Small models run ONNX on CPU and the VLM runs llama.cpp in Vulkan mode. NVIDIA GPU users can install mineru[full]>=4.0 for better throughput.

Which file types does MinerU not parse?

MinerU does not parse plain-text files such as .txt, .md, .markdown, .rst, and .tex. The guide says to read those directly.

Sources

MinerU README badge

Maintainers can show the tool's weekly Board rank in their README:

[![Devtoolwatch](https://index.devtoolwatch.com/badge/mineru.svg)](https://index.devtoolwatch.com/tools/mineru/)

How this page is made: Devtoolwatch writes each review from the project's README, releases, package metadata and website, with an AI model, then checks every claim against the source it cites and drops what the sources don't support. The numbers come from Devtoolwatch's daily GitHub snapshots, GitHub's stargazer history, and the npm and PyPI download APIs. Spotted an error? Email [email protected].