MinerU
Transforms complex documents like PDFs and Office docs into LLM-ready markdown/JSON for your Agentic workflows.
MinerU is an MCP server that converts complex documents—including PDFs, Office files, images, and more—into LLM-ready markdown or JSON for agentic workflows. It offers four quality tiers (flash, basic, standard, advanced) balancing speed and parsing accuracy, supporting OCR, tables, formulas, and multi-format inputs. Agents can progressively read documents page-by-page or block-by-block with stable citation locators, search indexed content, and retrieve visual block images. The server runs locally by default with configurable backends (ONNX, Torch, llama.cpp, vLLM, LMDeploy) and optionally supports remote parsing.
Key Features
Use Cases
- 01Reading and parsing academic papers with complex formulas and tables for research agents
- 02Converting scanned documents and images into structured text using OCR for archival workflows
- 03Progressive document analysis where agents read long reports page-by-page without loading entire files into context
- 04Building document search and citation systems with stable locators for legal or compliance agents
- 05Extracting structured data from invoices, contracts, and forms across multiple Office formats
- 06Creating local document libraries with automatic indexing and content search for knowledge management agents
Related MCP Servers
View moremarkitdown
Python tool for converting files and office documents to Markdown.
firecrawl
The context API to search, scrape, and interact with the web at scale. 🔥
prompts.chat
f.k.a. Awesome ChatGPT Prompts. Share, discover, and collect prompts from the community. Free and open source — self-host for your organization with complete privacy.
langflow
Langflow is a powerful tool for building and deploying AI-powered agents and workflows.
MinerU — FAQ
What is MinerU?+
MinerU is an MCP server that parses PDFs, Office documents, images, and other formats into markdown or JSON for AI agents. It provides four quality tiers, local document library management, and progressive reading with stable citation locators.
How do I install MinerU?+
Install MinerU 4.0+ using uv, pipx, or pip with Python >=3.10,<3.15. The recommended approach is 'uv tool install --python 3.12 "mineru>=4.0,<5"', then start the background service with 'mineru server start'.
Which AI clients work with MinerU?+
MinerU works with any MCP-compatible client including Claude Desktop and other clients supporting the Model Context Protocol. It provides a Python SDK and V1 API for custom integrations.
Do I need API keys or special hardware?+
No API keys are required for local parsing. The default install runs on CPU using ONNX and llama.cpp. For higher throughput, GPU acceleration with Torch, vLLM, or LMDeploy is optional but recommended for standard/advanced tiers.
Is MinerU free?+
Yes, MinerU is open-source under the MinerU Open Source License (based on Apache 2.0). All local parsing is free; optional remote parsing may have quota limits.
What quality tier should I use?+
Use standard tier (default) for normal reading. Use flash for quick previews and indexing, basic for lower-resource local parsing, and advanced only for difficult documents when you accept longer processing time.
How do I install MinerU?+
Open the source repository on GitHub and follow its README. MinerU is a mcp server — MCP Agents Market links you directly to the official repo.
Is MinerU free?+
MinerU is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.