MinerU

Transforms complex documents like PDFs into LLM-ready markdown/JSON for your Agentic workflows.

80.9k
Stars
+3758
Stars/month
911
Commits (90d)
10
Releases (6m)

Star Growth

+23.6k (41.0%)
56.2k69.4k82.6kMar 27Oct 1

Overview

MinerU 是一个专为 AI 代理工作流设计的文档处理工具,能够将复杂的 PDF 和其他文档格式转换为 LLM 可处理的 Markdown 和 JSON 格式。该工具通过智能解析技术,保持文档的结构完整性和内容准确性,特别适合需要大规模文档处理的 AI 应用场景。MinerU 提供了 Python SDK 和在线 Web 应用两种使用方式,满足不同开发者的需求。其强大的文档理解能力使得原本非结构化的 PDF 内容可以被 AI 模型有效利用,大大提升了文档驱动的 AI 应用的数据质量。凭借超过 57,000 个 GitHub 星标,MinerU 已成为文档预处理领域的热门选择,被广泛应用于知识库构建、RAG 系统和智能文档分析等场景。工具支持批量处理,能够高效处理大量文档,为企业级 AI 应用提供可靠的文档转换基础设施。

Deep Analysis

Key Differentiator

Unlike PyPDF (text-only extraction) or cloud OCR services, MinerU preserves document layout including tables with embedded formulas and achieves 86.2 on OmniDocBench — purpose-built for AI/RAG document pipelines

⚡ Capabilities

  • • High-precision PDF/document parsing with layout detection and formula recognition
  • • Native DOCX parsing with end-to-end speed improvement of 10x+ over PDF conversion
  • • Table parsing including images/formulas inside table cells
  • • Seal text recognition, vertical text support, and interline formula numbering
  • • 86.2 score on OmniDocBench benchmark for document understanding
  • • Web app, desktop client, and API access via mineru.net

🔗 Integrations

Hugging Face SpacesModelScopeGoogle Colab

✓ Best For

  • ✓ RAG pipelines needing high-fidelity PDF/DOCX extraction with tables, formulas, and layouts preserved
  • ✓ Academic and research teams processing scientific papers with complex mathematical notation

✗ Not Ideal For

  • ✗ Simple text extraction from clean documents — use PyPDF or pdfplumber instead
  • ✗ Real-time OCR on images — use Tesseract or cloud vision APIs

Languages

Python

Deployment

pip installWeb app (mineru.net)Desktop clientAPI accessDocker

Pricing Detail

Free: Open source (AGPL-like), free web demo on mineru.net
Paid: API access via mineru.net (usage-based)

⚠ Known Limitations

  • ⚠ GPU recommended for fast processing — CPU-only mode is significantly slower
  • ⚠ Focused on document extraction only — not a general-purpose OCR tool
  • ⚠ AGPL-style license may restrict commercial use

Pros

  • + 专门针对 LLM 优化的输出格式,确保转换后的 Markdown/JSON 能够被 AI 模型高质量理解和处理
  • + 支持复杂 PDF 文档的结构化解析,保持表格、图像和文本布局的完整性
  • + 提供 Python SDK 和 Web 应用双重接口,既适合程序化集成也支持交互式使用

Cons

  • - 主要专注于 PDF 处理,对其他文档格式的支持可能有限
  • - 复杂文档的处理质量可能依赖于原始文档的质量和结构清晰度
  • - 大规模批量处理时可能需要考虑计算资源和处理时间的平衡

Use Cases

  • • 构建 RAG(检索增强生成)系统时,将企业内部 PDF 文档转换为向量数据库可索引的格式
  • • 为 AI 代理开发智能文档分析功能,自动提取和结构化合同、报告中的关键信息
  • • 建立知识管理系统,将历史文档资料转换为可搜索和可查询的结构化数据

Getting Started

1. 通过 pip 安装 MinerU:`pip install mineru` 2. 访问 mineru.net 体验在线版本或在 Python 中导入库开始文档转换 3. 使用简单的 API 调用处理第一个 PDF 文件,查看生成的 Markdown 输出效果

Alternatives

See all 8 MinerU alternatives →

Works with MinerU

Tools that integrate with MinerU, often used together in the same stack.

Compare MinerU

Maintain MinerU?

Show your live rank in your README, or put MinerU in front of every visitor to AgentoolRank.