论文、合同、扫描件丢给 AI 之前，最难的一步往往是「先把 PDF 洗干净」。这几个开源项目专干这件事：转成 Markdown/JSON，直接喂给 RAG 或 agent。

① MarkItDown · 微软出品，Office/PDF/图片一键转 Markdown， 15万+，格式覆盖最全 https://github.com/microsoft/markitdown…

② MinerU · 复杂 PDF/Office 转 LLM 友好 Markdown/JSON， 6.8万，论文和研报解析很能打 https://github.com/opendatalab/MinerU…

③ Docling · IBM 开源，文档清洗后接 gen AI pipeline， 6.1万，企业文档场景友好 https://github.com/docling-project/docling…

④ marker · 高精度 PDF → Markdown + JSON， 3.6万、GPL-3.0，扫描版也能救 https://github.com/datalab-to/marker…

⑤ surya · OCR + 版面分析 + 表格识别，90+ 语言， 2万、Apache-2.0，复杂排版先过它 https://github.com/datalab-to/surya…

怎么选 日常 Office/PDF 快速转文本 → MarkItDown 论文/研报/复杂版式 → MinerU 或 marker 接企业 RAG 流水线 → Docling 扫描件、表格多、排版乱 → surya 预处理 + MinerU

microsoft/markitdown

Source: https://github.com/microsoft/markitdown

MarkItDown

MarkItDown performs I/O with the privileges of the current process. Like open() or requests.get(), it will access resources that the process itself can access. Sanitize your inputs in untrusted environments, and call the narrowest convert_* function needed for your use case (e.g., convert_stream(), or convert_local()). See the Security Considerations section of the documentation for more information.

MarkItDown is a lightweight Python utility for converting various files to Markdown for use with LLMs and related text analysis pipelines. To this end, it is most comparable to textract, but with a focus on preserving important document structure and content as Markdown (including: headings, lists, tables, links, etc.) While the output is often reasonably presentable and human-friendly, it is meant to be consumed by text analysis tools – and may not be the best option for high-fidelity document conversions for human consumption.

MarkItDown currently supports the conversion from:

• PDF
• PowerPoint
• Word
• Excel
• Images (EXIF metadata and OCR)
• Audio (EXIF metadata and speech transcription)
• HTML
• Text-based formats (CSV, JSON, XML)
• ZIP files (iterates over contents)
• Youtube URLs
• EPubs
• … and more!

Why Markdown?

Markdown is extremely close to plain text, with minimal markup or formatting, but still provides a way to represent important document structure. Mainstream LLMs, such as OpenAI’s GPT-4o, natively “speak” Markdown, and often incorporate Markdown into their responses unprompted. This suggests that they have been trained on vast amounts of Markdown-formatted text, and understand it well. As a side benefit, Markdown conventions are also highly token-efficient.

Prerequisites

MarkItDown requires Python 3.10 or higher. It is recommended to use a virtual environment to avoid dependency conflicts.

With the standard Python installation, you can create and activate a virtual environment using the following commands:

python -m venv .venv
source .venv/bin/activate

If using uv, you can create a virtual environment with:

uv venv --python=3.12 .venv
source .venv/bin/activate
# NOTE: Be sure to use 'uv pip install' rather than just 'pip install' to install packages in this virtual environment

If you are using Anaconda, you can create a virtual environment with:

conda create -n markitdown python=3.12
conda activate markitdown

Installation

To install MarkItDown, use pip: pip install 'markitdown[all]'. Alternatively, you can install it from the source:

git clone [email protected]:microsoft/markitdown.git
cd markitdown
pip install -e 'packages/markitdown[all]'

Usage

Command-Line

markitdown path-to-file.pdf > document.md

Or use -o to specify the output file:

markitdown path-to-file.pdf -o document.md

You can also pipe content:

cat path-to-file.pdf | markitdown

Optional Dependencies

MarkItDown has optional dependencies for activating various file formats. Earlier in this document, we installed all optional dependencies with the [all] option. However, you can also install them individually for more control. For example:

pip install 'markitdown[pdf, docx, pptx]'

will install only the dependencies for PDF, DOCX, and PPTX files.

At the moment, the following optional dependencies are available:

• [all] Installs all optional dependencies
• [pptx] Installs dependencies for PowerPoint files
• [docx] Installs dependencies for Word files
• [xlsx] Installs dependencies for Excel files
• [xls] Installs dependencies for older Excel files
• [pdf] Installs dependencies for PDF files
• [outlook] Installs dependencies for Outlook messages
• [az-doc-intel] Installs dependencies for Azure Document Intelligence
• [az-content-understanding] Installs dependencies for Azure Content Understanding
• [audio-transcription] Installs dependencies for audio transcription of wav and mp3 files
• [youtube-transcription] Installs dependencies for fetching YouTube video transcription

Plugins

MarkItDown also supports 3rd-party plugins. Plugins are disabled by default. To list installed plugins:

markitdown --list-plugins

To enable plugins use:

markitdown --use-plugins path-to-file.pdf

To find available plugins, search GitHub for the hashtag #markitdown-plugin. To develop a plugin, see packages/markitdown-sample-plugin.

markitdown-ocr Plugin

The markitdown-ocr plugin adds OCR support to PDF, DOCX, PPTX, and XLSX converters, extracting text from embedded images using LLM Vision — the same llm_client / llm_model pattern that MarkItDown already uses for image descriptions. No new ML libraries or binary dependencies required.

Installation:

pip install markitdown-ocr
pip install openai  # or any OpenAI-compatible client

Usage:

Pass the same llm_client and llm_model you would use for image descriptions:

from markitdown import MarkItDown
from openai import OpenAI

md = MarkItDown(
    enable_plugins=True,
    llm_client=OpenAI(),
    llm_model="gpt-4o",
)
result = md.convert("document_with_images.pdf")
print(result.text_content)

If no llm_client is provided the plugin still loads, but OCR is silently skipped and the standard built-in converter is used instead.

See packages/markitdown-ocr/README.md for detailed documentation.

Azure Content Understanding

Azure Content Understanding provides higher-quality conversion with structured field extraction (YAML front matter), multi-modal support (documents, images, audio, video), and configurable analyzers.

Install: pip install 'markitdown[az-content-understanding]'

When to use Content Understanding

Content Understanding is ideal when you need capabilities beyond what built-in or Document Intelligence converters provide:

• Audio and video files — CU is the only option for video, and the higher-quality cloud option for audio. Built-in converters have no video support and only basic audio transcription.
• Structured field extraction — Prebuilt or custom-built analyzers extract domain-specific fields (invoice amounts, receipt dates, contract clauses) serialized as YAML front matter. Neither built-in nor Doc Intel integration exposes fields.
• Higher-quality document extraction — Cloud-based layout analysis and OCR for scanned PDFs, complex tables, and multi-page documents.
• Single API for all modalities — One cu_endpoint handles documents, images, audio, and video with automatic analyzer routing.

CLI:

markitdown path-to-file.pdf --use-cu --cu-endpoint "<content_understanding_endpoint>"

Python API:

from markitdown import MarkItDown

# Zero-config — auto-selects analyzer per file type
md = MarkItDown(cu_endpoint="<content_understanding_endpoint>")
result = md.convert("report.pdf")   # documents → prebuilt-documentSearch
result = md.convert("meeting.mp4")  # video → prebuilt-videoSearch
result = md.convert("call.wav")     # audio → prebuilt-audioSearch
print(result.markdown)

With a custom analyzer (for domain-specific field extraction):

md = MarkItDown(
    cu_endpoint="<content_understanding_endpoint>",
    cu_analyzer_id="my-invoice-analyzer",
)
result = md.convert("invoice.pdf")
print(result.markdown)
# Output includes YAML front matter with extracted fields:
# ---
# contentType: document
# fields:
#   VendorName: CONTOSO LTD.
#   InvoiceDate: '2019-11-15'
# ---
# <!-- page 1 -->
# ...

When cu_analyzer_id is set, the converter automatically scopes it to compatible file types based on the analyzer’s modality. Incompatible types (e.g., audio files with a document analyzer) auto-route to default prebuilt analyzers.

Cost note: Each convert() call for a CU-routed format is a billable Azure API call. Use cu_file_types to restrict which formats route to CU:

from markitdown.converters import ContentUnderstandingFileType

md = MarkItDown(
    cu_endpoint="<content_understanding_endpoint>",
    cu_file_types=[ContentUnderstandingFileType.PDF],  # only PDFs use CU
)

More information about Azure Content Understanding can be found here.

Azure Document Intelligence

To use Microsoft Document Intelligence for conversion:

markitdown path-to-file.pdf -o document.md -d -e "<document_intelligence_endpoint>"

More information about how to set up an Azure Document Intelligence Resource can be found here

Python API

Basic usage in Python:

from markitdown import MarkItDown

md = MarkItDown(enable_plugins=False) # Set to True to enable plugins
result = md.convert("test.xlsx")
print(result.text_content)

Document Intelligence conversion in Python:

from markitdown import MarkItDown

md = MarkItDown(docintel_endpoint="<document_intelligence_endpoint>")
result = md.convert("test.pdf")
print(result.text_content)

To use Large Language Models for image descriptions (currently only for pptx and image files), provide llm_client and llm_model:

from markitdown import MarkItDown
from openai import OpenAI

client = OpenAI()
md = MarkItDown(llm_client=client, llm_model="gpt-4o", llm_prompt="optional custom prompt")
result = md.convert("example.jpg")
print(result.text_content)

Docker

docker build -t markitdown:latest .
docker run --rm -i markitdown:latest < ~/your-file.pdf > output.md

Contributing

This project welcomes contributions and suggestions. Most contributions require you to agree to a Contributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us the rights to use your contribution. For details, visit https://cla.opensource.microsoft.com.

When you submit a pull request, a CLA bot will automatically determine whether you need to provide a CLA and decorate the PR appropriately (e.g., status check, comment). Simply follow the instructions provided by the bot. You will only need to do this once across all repos using our CLA.

This project has adopted the Microsoft Open Source Code of Conduct. For more information see the Code of Conduct FAQ or contact [email protected] with any additional questions or comments.

How to Contribute

You can help by looking at issues or helping review PRs. Any issue or PR is welcome, but we have also marked some as ‘open for contribution’ and ‘open for reviewing’ to help facilitate community contributions. These are of course just suggestions and you are welcome to contribute in any way you like.

Running Tests and Checks

• Navigate to the MarkItDown package:

  cd packages/markitdown
  

• Install hatch in your environment and run tests:

  pip install hatch  # Other ways of installing hatch: https://hatch.pypa.io/dev/install/
  hatch shell
  hatch test
  

  (Alternative) Use the Devcontainer which has all the dependencies installed:

  # Reopen the project in Devcontainer and run:
  hatch test
  

• Run pre-commit checks before submitting a PR: pre-commit run --all-files

Security Considerations

MarkItDown performs I/O with the privileges of the current process. Like open() or requests.get(), it will access resources that the process itself can access.

Sanitize your inputs: Do not pass untrusted input directly to MarkItDown. If any part of the input may be controlled by an untrusted user or system, such as in hosted or server-side applications, it must be validated and restricted before calling MarkItDown. Depending on your environment, this may include restricting file paths, limiting URI schemes and network destinations, and blocking access to private, loopback, link-local, or metadata-service addresses.

Call only the conversion method you need: Prefer the narrowest conversion API that fits your use case. MarkItDown’s convert() method is intentionally permissive and can handle local files, remote URIs, and byte streams. If your application only needs to read local files, call convert_local() instead. If you need more control over URI fetching, call requests.get() yourself and pass the response object to convert_response(). For maximum control, open a stream to the input you want converted and call convert_stream().

Contributing 3rd-party Plugins

You can also contribute by creating and sharing 3rd party plugins. See packages/markitdown-sample-plugin for more details.

Trademarks

This project may contain trademarks or logos for projects, products, or services. Authorized use of Microsoft trademarks or logos is subject to and must follow Microsoft’s Trademark & Brand Guidelines. Use of Microsoft trademarks or logos in modified versions of this project must not cause confusion or imply Microsoft sponsorship. Any use of third-party trademarks or logos are subject to those third-party’s policies.