Skip to main content

mddocx

PyPI Python CI Release License: MIT

Deterministic Markdown-to-DOCX compilation for professional, editable Microsoft Word documents.

mddocx converts Markdown into native Word structures instead of screenshots or flattened pages. Headings, lists, tables, equations, charts, links, references, footnotes, headers, footers, and other document elements remain editable in Microsoft Word.

What it does

  • Converts Markdown files or strings into native DOCX documents.
  • Preserves editable Word headings, paragraphs, lists, tables, hyperlinks, and code blocks.
  • Supports nested lists, task lists, blockquotes, callouts, page breaks, and section breaks.
  • Produces editable Word equations using OMML rather than images.
  • Creates editable Office charts backed by embedded Excel workbooks.
  • Supports CSV and JSON data for charts and native Word tables.
  • Handles local, HTTPS, and Base64 data:image/...;base64,... images, plus captions, bookmarks, cross-references, footnotes, endnotes, and citations.
  • Supports themes, fonts, templates, right-to-left text, headers, footers, tables of contents, and page fields.
  • Provides project mode for compiling multiple Markdown chapters into one document.
  • Includes diagnostics, document inspection, accessibility checks, and profiling utilities.
  • Provides privacy-aware metadata handling for exported AI/chat conversations.
  • Accepts common equation syntax copied from ChatGPT, Claude, Gemini, and other MathJax/LaTeX-producing assistants.

Installation

mddocx requires Python 3.11 or newer.

Install from PyPI

The distribution is published as mddocx-native; the Python import and command remain mddocx:

python -m pip install --upgrade mddocx-native

Verify the installation:

mddocx --version
mddocx doctor

Install a downloaded wheel

Download the .whl file from the latest GitHub Release, open a terminal in the download directory, and run:

python -m pip install ./mddocx_native-1.2.4-py3-none-any.whl

On Windows PowerShell, the equivalent command is:

python -m pip install .\mddocx_native-1.2.4-py3-none-any.whl

You can also install the wheel directly from the v1.2.4 GitHub Release:

python -m pip install "https://github.com/TasinAhmed2508/mddocx/releases/download/v1.2.4/mddocx_native-1.2.4-py3-none-any.whl"

Install from GitHub

python -m pip install "git+https://github.com/TasinAhmed2508/mddocx.git@v1.2.4"

Install for local development

git clone https://github.com/TasinAhmed2508/mddocx.git
cd mddocx
python -m pip install -e ".[dev]"

Optional features:

python -m pip install "mddocx-native[math]"
python -m pip install "mddocx-native[images]"
python -m pip install "mddocx-native[math,images]"

Quick start

Create a file named report.md:

# Quarterly Report

## Summary

This report is generated from Markdown and remains fully editable in Word.

| Metric | Value |
| --- | ---: |
| Revenue | 125,000 |
| Growth | 18% |

Convert it to Word:

mddocx report.md -o report.docx

Command-line interface

mddocx document.md -o document.docx
mddocx document.md --theme academic
mddocx document.md --template company.docx
mddocx document.md --toc --page-numbers
mddocx document.md --header "Project Report" --footer "Confidential"
mddocx document.md --diagnostics-json diagnostics.json
mddocx document.md --strict-math
mddocx math-check document.md
mddocx inspect document.docx --strict
mddocx accessibility document.docx --strict
mddocx doctor
mddocx api --json

For multi-file document projects:

mddocx project init report
mddocx build report
mddocx project info report
mddocx watch report

For interactive use:

mddocx shell

Run mddocx --help for the complete command reference.

Python API

from mddocx import RenderConfig, render, render_string

render("report.md", "report.docx")

config = RenderConfig(theme="modern")
document = render_string("# Generated report\n\nEditable Word content.", config=config)

with open("generated.docx", "wb") as file:
    file.write(document)

The public API is documented in docs/API_STABILITY.md and can also be inspected with mddocx api --json. The additive compiler-stage mapping and behavioral changes are documented in docs/V2_MIGRATION.md.

For callers that need the output, diagnostics, and profiling data together, use the staged-v2 compiler service:

from mddocx import Compiler

result = Compiler().compile_string("# Report\n\nInline \\(x^2\\).")
print(result.success, result.diagnostics, result.stats)
print(result.layout_plan.tables)

The same service exposes parse_string, parse_file, normalize, plan, render, check_string, check_file, compile_string, and compile_file, so applications can inspect canonical AST and layout decisions without duplicating compiler internals.

AI-generated equations

Common inline forms (\\(...\\) and $...$), display forms (\\[...\\] and $$...$$), and math, latex, or tex fenced blocks are converted to editable Word equations. Math-like text inside code spans and ordinary code fences is preserved as code, and common currency forms are preserved as text.

When an equation uses syntax that the active math engine cannot convert, mddocx preserves its source as editable text and emits a MATH201 warning. Use --strict-math when every equation must be native Word math and conversion should fail on the first unsupported expression. See docs/AI_MATH_COMPATIBILITY.md for the tested compatibility contract.

Run mddocx math-check document.md before conversion to see the active math engine and the number of equations that can be rendered natively. Add --json for automation or --strict to return exit code 3 when any equation requires a fallback.

YAML front matter

Document settings can be defined at the top of a Markdown file:

---
title: Research Report
author: Example Author
theme: academic
toc: true
page_numbers: true
auto_landscape_tables: true
header: Research Group
footer: Confidential
---

Privacy and security defaults

  • Safe public HTTPS resources are allowed by default. The interactive shell asks on first launch whether they should instead be blocked with RESOURCE201, and remembers that preference.
  • Base64 image data URIs are decoded locally, never sent over the network, and remain subject to the configured per-resource size limit and image validation.
  • HTTPS image downloads can be explicitly enabled and restricted by domain.
  • Use --block-remote-resources for a single noninteractive conversion, or --allow-remote-resources to override a saved shell block for one render.
  • Local resource paths cannot escape the document or project directory.
  • Download size, MIME type, redirects, and network destinations are constrained.
  • SVG parsing rejects external references and uses hardened XML handling.
  • Code blocks and TeX are never executed.
  • AI/chat export metadata is removed conservatively before Markdown parsing when enabled.

Project layout

mddocx/
├── src/mddocx/        # Package source code
├── docs/              # User and technical documentation
├── dist/              # Local build output, when generated
├── .github/workflows/ # CI and release automation
├── pyproject.toml     # Package metadata and build configuration
└── README.md          # Project overview and usage guide

Documentation

Detailed documentation is available in the docs directory, including compatibility, API stability, interactive CLI, charts and data, AI export metadata, accessibility, and extensions.

See the 1.2.4 release notes for the Base64-image, AI-math, compiler, layout, security-policy, and migration summary.

The compiler stages and system architecture are described in docs/ARCHITECTURE.md. The machine-readable supported feature contract is docs/feature-matrix.json.

The machine-readable fidelity contract is docs/feature-matrix.json; every listed feature links its accepted syntax, canonical AST, native Word representation, diagnostic behavior, host status, and tracked regression evidence.

The complete staged engineering program is preserved in docs/FIDELITY_FIRST_V2_PLAN.md; progress and compatibility notes remain additive until a qualified 2.x release is ready.

Microsoft Word 365 release qualification, including the reusable Windows automation script, is documented in docs/WORD365_QA.md.

Releases

Release builds are generated by GitHub Actions. Each versioned release publishes the wheel and source archive as downloadable GitHub Release assets.

License

Released under the MIT License.

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

mddocx_native-1.2.4.tar.gz (165.2 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

mddocx_native-1.2.4-py3-none-any.whl (176.0 kB view details)

Uploaded Python 3

File details

Details for the file mddocx_native-1.2.4.tar.gz.

File metadata

  • Download URL: mddocx_native-1.2.4.tar.gz
  • Upload date:
  • Size: 165.2 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for mddocx_native-1.2.4.tar.gz
Algorithm Hash digest
SHA256 ab37bf5aead47979bd8f76094f30e65d869e79e907640cef23f1388d71b33fbd
MD5 a511b30d008a09b0852a94a78ec9fde7
BLAKE2b-256 25928e6e67581a618258b694e4795c4b1d87bcc84a6983daf0f129ef785ebb11

See more details on using hashes here.

Provenance

The following attestation bundles were made for mddocx_native-1.2.4.tar.gz:

Publisher: release.yml on TasinAhmed2508/mddocx

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file mddocx_native-1.2.4-py3-none-any.whl.

File metadata

  • Download URL: mddocx_native-1.2.4-py3-none-any.whl
  • Upload date:
  • Size: 176.0 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for mddocx_native-1.2.4-py3-none-any.whl
Algorithm Hash digest
SHA256 34fc68fa9b7985864320c34cfdcb77aa51097d291ca2bdf6470251e13d4aeaa4
MD5 208204281b54d38df9e7e6c99bafcf3e
BLAKE2b-256 1f3d165f6e112b309ad29a8235797402575593169528438a48d4f51b33220c26

See more details on using hashes here.

Provenance

The following attestation bundles were made for mddocx_native-1.2.4-py3-none-any.whl:

Publisher: release.yml on TasinAhmed2508/mddocx

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Release history Release notifications | RSS feed

This release

1.2.4 This release

2 files

1.2.3

2 files

1.2.2

2 files

1.2.1

2 files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page