HTML FOR DOCX

Convert html to docx, this project is a fork from the discontinued pqzx/html2docx.

How to install

pip install html-for-docx

Usage

The basic usage

Add HTML-formatted content to an existing .docx document

from docx import Document
from html4docx import HtmlToDocx

document = Document('existing_file.docx')
parser = HtmlToDocx()
html_string = '<h1>Hello world</h1>'
parser.add_html_to_document(html_string, document)
document.save('existing_file.docx')

You can use python-docx to manipulate directly the file, here an example

from docx import Document
from html4docx import HtmlToDocx

document = Document()
parser = HtmlToDocx()

html_string = '<h1>Hello world</h1>'
parser.add_html_to_document(html_string, document)

document.save('your_file_name.docx')

or incrementally add new html to document and save it when finished, new content will always be added at the end

from docx import Document
from html4docx import HtmlToDocx

document = Document()
parser = HtmlToDocx()

for part in ['First', 'Second', 'Third']:
    parser.add_html_to_document(f'<h1>{part} Part</h1>', document)

parser.save('your_file_name.docx')

When you pass a Document object, you can either use document.save() from python-docx or parser.save() from html4docx, both work well.

Both support saving it in-memory, using BytesIO.

from io import BytesIO
from docx import Document
from html4docx import HtmlToDocx

buffer = BytesIO()
document = Document()
parser = HtmlToDocx()

html_string = '<h1>Hello world</h1>'
parser.add_html_to_document(html_string, document)

# Save the document to the in-memory buffer
parser.save(buffer)

# If you need to read from the buffer again after saving,
# you might need to reset its position to the beginning
buffer.seek(0)

Convert files directly

from html4docx import HtmlToDocx

parser = HtmlToDocx()
parser.parse_html_file(input_html_file_path, output_docx_file_path)
# You can also define a encoding, by default is utf-8
parser.parse_html_file(input_html_file_path, output_docx_file_path, 'utf-8')

If the HTML file references local images or stylesheets, allow the folder that holds them (see External resources security):

import os

parser.options['local-files-root'] = os.path.dirname(os.path.abspath(input_html_file_path))
parser.parse_html_file(input_html_file_path, output_docx_file_path)

Convert files from a string

from html4docx import HtmlToDocx

parser = HtmlToDocx()
docx = parser.parse_html_string(input_html_file_string)

Change table styles

Tables are not styled by default. Use the table_style attribute on the parser to set a table style before convert html. The style is used for all tables.

from html4docx import HtmlToDocx

parser = HtmlToDocx()
parser.table_style = 'Light Shading Accent 4'
docx = parser.parse_html_string(input_html_file_string)

To add borders to tables, use the Table Grid style:

parser.table_style = 'Table Grid'

All table styles we support can be found here.

Options

All options are set through parser.options before parsing. Copy only the ones you need to change:

from html4docx import HtmlToDocx

parser = HtmlToDocx()

# --- Rendering ---
parser.options['images'] = False         # Default True.  False ignores all images.
parser.options['tables'] = False         # Default True.  False renders only the raw table content.
parser.options['styles'] = False         # Default True.  False ignores all CSS (also disables 'style-map').
parser.options['fix-html'] = False       # Default True.  False skips BeautifulSoup fixing of missing HTML tags.
parser.options['style-map'] = False      # Default True.  False ignores CSS class to Word style mapping.
parser.options['tag-override'] = False   # Default True.  False ignores HTML tag overrides.
parser.options['html-comments'] = True   # Default False. True renders "<!-- ... -->" comments.

# --- External resources security (see the section below) ---
parser.options['allow-remote-resources'] = False                    # Default True.  False skips all http(s) images and stylesheets.
parser.options['private-network-allowlist'] = ["10.0.5.20", "192.168.10.0/24"]  # Default []. Private IPs/CIDRs that may be fetched.
parser.options['local-files-root'] = '/srv/app/assets'              # Default None.  Directory local files may be read from (None = disabled).
parser.options['max-resource-bytes'] = 5 * 1024 * 1024              # Default 10 MB. Maximum size of each fetched resource.
parser.options['allow-private-network'] = False                     # Default False. True allows EVERY private address and ignores the allowlist: trusted HTML only!

docx = parser.parse_html_string(input_html_file_string)

External resources security

HTML is often untrusted input, so images (<img src>) and stylesheets (<link rel="stylesheet">) are fetched under a restrictive policy by default:

• Remote resources: only http/https are fetched. file://, ftp:// and any other scheme are always refused. Inline images (data:image/...) keep working.
• Private network: requests to non-public addresses (loopback, private ranges, link-local such as the cloud metadata endpoint 169.254.169.254, reserved, multicast) are refused. The check is applied to the IP actually connected to, on every redirect hop, so DNS rebinding and open redirects can't bypass it.
• Local files: disabled by default. When local-files-root is set, only files inside that directory (after resolving .. and symlinks) can be read, and relative paths are resolved against it.
• Size: each resource, remote or local, is limited to max-resource-bytes.

Reaching an internal server. Add its addresses to private-network-allowlist. Only those IPs/CIDRs are opened, and everything else stays blocked. Public addresses (CDNs, websites) never need to be listed. Keep entries as narrow as possible: a broad range such as 169.254.0.0/16 also opens cloud metadata endpoints. Invalid entries raise ValueError instead of being silently ignored. allow-private-network opens every private address at once and should only be used when the HTML is fully trusted.

Converting HTML from users? Keep allow-private-network disabled, keep private-network-allowlist as small as possible and never point local-files-root at a directory containing sensitive files. If you don't need external images or stylesheets at all, set allow-remote-resources to False.

Blocked or unavailable resources are skipped and a warning is logged (logger html4docx.resources); the conversion itself does not fail.

Migrating from <=1.2.x

Public http/https URLs and data:image/... images work as before. What changed:

• Local paths in <img src> and <link href> are disabled by default. Set local-files-root to the folder that holds your assets. Relative paths are now resolved against it, not the current working directory:

import os

parser.options['local-files-root'] = os.path.abspath('./assets')

• Private network addresses (e.g. an intranet server) are blocked. Add them to private-network-allowlist.
• Proxies: HTTP_PROXY / HTTPS_PROXY environment variables are ignored, because routing through a proxy would bypass the private network protection. If your server can only reach the internet through a proxy, remote resources will not load. Embed images as data:image/... URIs, put CSS in an inline <style> tag, or serve the files from local-files-root.

Extended Styling Features

CSS Class to Word Style Mapping

Map HTML CSS classes to Word document styles:

from html4docx import HtmlToDocx

style_map = {
    'code-block': 'Code Block',
    'numbered-heading-1': 'Heading 1 Numbered',
    'finding-critical': 'Finding Critical'
}

parser = HtmlToDocx(style_map=style_map)
parser.add_html_to_document(html, document)

Tag Style Overrides

Override default tag-to-style mappings:

tag_overrides = {
    'h1': 'Custom Heading 1',  # All <h1> use this style
    'pre': 'Code Block'        # All <pre> use this style
}

parser = HtmlToDocx(tag_style_overrides=tag_overrides)

Custom styles from a Word template: Use a document created from a .docx that already defines the styles (e.g. "Code Block", "Custom Markdown"). Pass that same document to the parser and save it so the custom styles are preserved:

from docx import Document
from html4docx import HtmlToDocx

doc = Document("path/to/template.docx")  # template has Code Block, Custom Markdown, etc.
parser = HtmlToDocx(tag_style_overrides={"code": "Custom Markdown", "pre": "Code Block"})
parser.add_html_to_document(html, doc)
doc.save("output.docx")  # save the template-based doc so custom styles are preserved

If you save a different document (for example, by creating a new Document() instead of loading your template), the output file will not contain the template’s custom styles.

If a referenced custom style does not exist in the document at generation time, a warning will be logged to help you detect the missing style.

Default Paragraph Style

Set custom default paragraph style:

# Word's 'Normal' style is used by default
parser = HtmlToDocx()

# Use a custom style, e.g. 'Body', as default
parser = HtmlToDocx(default_paragraph_style='Body')

CSS

Inline CSS Styles

Full support for inline CSS styles on any element:

<p style="color: red; font-size: 14pt">Red 14pt paragraph</p>
<span style="font-weight: bold; color: blue">Bold blue text</span>

External CSS Styles

Limited support for external CSS styles (local or url). Remote stylesheets follow the External resources security rules, and local ones require local-files-root:

HTML

<head>
    <title>My page</title>
    <!-- Must contain rel="stylesheet" otherwise will be ignored -->
    <link rel="stylesheet" href="https://my-external-css.domain.com/style.css">
    <!-- OR, a file inside local-files-root -->
    <link rel="stylesheet" href="css/style.css">
</head>
<body>
    <div class="classFromExternalCss">
        <p>This should be Red and <span class="blueSpan">Blue</span>.<p>
        <p id="specialParagraph" style="font-weight: normal">A very special paragraph.</p>
    </div>
</body>

CSS

#specialParagraph {
  text-decoration: underline solid red;
  color: gray
}

.blueSpan {
  color: blue;
}

p {
  color: red;
  font-weight: bold;
}

Supported CSS properties:

• color
• font-size
• font-weight (bold)
• font-style (italic)
• text-decoration (underline, line-through)
• font-family
• text-align
• background-color
• Border (for tables)
• Vertical Align (for tables)

!important Flag Support

Proper CSS precedence with !important:

<span style="color: gray">
  Gray text with <span style="color: red !important">red important</span>.
</span>

The !important flag ensures highest priority.

Style Precedence Order

Styles are applied in this order (lowest to highest priority):

1. Base HTML tag styles (<b>, <em>, <code>)
2. Parent span styles
3. CSS class-based styles (from style_map)
4. External CSS Styles (from style tag or external files)
5. Inline CSS styles (from style attribute)
6. !important inline CSS styles (highest priority)

Metadata

You're able to read or set docx metadata:

from docx import Document
from html4docx import HtmlToDocx

document = Document()
parser = HtmlToDocx()
parser.set_initial_attrs(document)
metadata = parser.metadata

# Set new metadata
metadata.set_metadata(author="Jane", created="2025-07-18T09:30:00")

# You can get metadata as dict
metadata_json = metadata.get_metadata()
print(metadata_json['author']) # Jane
# or just print all metadata if you want
metadata.get_metadata(print_result=True)

document.save('your_file_name.docx')

You can find all available metadata attributes here.

Logging

html4docx uses Python's standard logging module with named, hierarchical loggers (for example, html4docx.h4d and html4docx.resources). The library never logs directly to the root logger and installs a NullHandler by default, so it remains silent unless your application configures logging.

Silence all html4docx logs:

import logging

logging.getLogger("html4docx").setLevel(logging.ERROR) # suppresses WARNING and below errors

Enable debug logging:

import logging

logging.basicConfig(level=logging.DEBUG)
logging.getLogger("html4docx").setLevel(logging.DEBUG)

Django / framework LOGGING dict — add an entry for the html4docx parent and it applies to all sub-loggers:

LOGGING = {
    "version": 1,
    "loggers": {
        "html4docx": {
            "level": "ERROR",   # suppresses WARNING and below
            "propagate": False,
        },
    },
}

Note: Unrecognised CSS properties (e.g. letter-spacing, margin, padding) are intentionally logged at DEBUG level because they are expected skips for any real-world HTML, not problems. Only genuinely unexpected situations (missing styles, unsupported colour formats, blocked external resources, etc.) are logged at WARNING.

Why

My goal in forking and fixing/updating this package was to complete my current task at work, which involves converting HTML to DOCX. The original package lacked a few features and had some bugs, preventing me from completing the task. Instead of creating a new package from scratch, I preferred to update this one.

Differences (fixes and new features)

Fixes

• Fix table_style not working | Dfop02 from Issue
• Handle missing run for leading br tag | dashingdove from PR
• Fix base64 images | djplaner from Issue
• Handle img tag without src attribute | johnjor from PR
• Fix bug when any style has !important | Dfop02
• Fix 'style lookup by style_id is deprecated.' | Dfop02
• Fix background-color not working | Dfop02
• Fix crashes when img or bookmark is created without paragraph | Dfop02
• Fix Ordered and Unordered Lists | TaylorN15 from PR
• Fixed styles was only being applied to span tag. | Dfop02 from Issue
• Fixed bug on styles parsing when style contains multiple colon. | Dfop02
• Fixed highlighting a single word | Lynuxen
• Fix color parsing failing due to invalid colors, falling back to black. | dfop02 from Issue
• Fix logging noise: replace root-logger calls with named module loggers so consumers can silence or configure html4docx output independently. | dfop02 from Issue
• Fixed non-breaking spaces ( ) being collapsed into ordinary spaces, which let Word break lines inside amounts like 1 000,00 € | Lynuxen
• Fix SSRF and arbitrary local file read through <img src> and <link rel="stylesheet"> when converting untrusted HTML | dfop02

New Features

• Add Width/Height style to images | maifeeulasad from PR
• Support px, cm, pt, in, rem, em, mm, pc and % units for styles | Dfop02
• Improve performance on large tables | dashingdove from PR
• Support for HTML Pagination | Evilran from PR
• Support Table style | Evilran from PR
• Support alternative encoding | HebaElwazzan from PR
• Support colors by name | Dfop02
• Support font_size when text, ex.: small, medium, etc. | Dfop02
• Support to internal links (Anchor) | Dfop02
• Support to rowspan and colspan in tables. | Dfop02 from Issue
• Support to 'vertical-align' in table cells. | Dfop02
• Support to metadata | Dfop02
• Add support to table cells style (border, background-color, width, height, margin) | Dfop02
• Being able to use inline images on same paragraph. | Dfop02
• Refactor Tests to be more consistent and less 'human validation' | Dfop02
• Support for common CSS properties for text | Lynuxen
• Support for CSS classes to Word Styles | raithedavion
• Support for HTML tag style overrides | raithedavion
• Support style tag and external CSS styles | Dfop02
• Security policy for external resources (allow-remote-resources, private-network-allowlist, allow-private-network, local-files-root, max-resource-bytes) | dfop02

To-Do

These are the ideas I'm planning to work on in the future to make this project even better:

• No ideas for now.

Known Limitations

Lists

• Maximum Nesting Depth: Ordered lists support up to 3 nested levels. Any additional depth beyond level 3 will be treated as level 3.
• Counter Reset Behavior:
  • At level 1, starting a new ordered list will reset the counter.
  • At levels 2 and 3, the counter will continue from the previous item unless explicitly reset.

CSS

1. Complex Selectors:

   • Do not support complex selector like div > p, h1 + h2.
   • Complex selectors are simplified to root tag.

2. Pseudo-classes:

   • Do not support :hover, :first-child, etc.
   • They're all removed during CSS Parsing

3. Media Queries:

   • Do not support @media queries.
   • Would be necessary a responsive CSS, that's not applied to Docx.

4. @import:

   • Do not support @import in CSS.
   • We can try implement it in future if necessary.

5. CSS Variables:

   • Do not support var(--name) values. Frameworks that define colors through variables (e.g. Bootstrap 5 text-danger) fall back to black.

External resources

• Proxies configured through HTTP_PROXY / HTTPS_PROXY are not used (see Migrating from <=1.2.x).
• data: URIs are only supported for images. Stylesheets must be remote, local (inside local-files-root) or inline in a <style> tag.

Project Guidelines

This project is primarily designed for compatibility with Microsoft Word, but it currently works well with LibreOffice and Google Docs, based on our testing. The goal is to maintain this cross-platform harmony while continuing to implement fixes and updates.

⚠️ However, please note that Microsoft Word is the priority. Bugs or issues specific to other editors (e.g., LibreOffice or Google Docs) may be considered, but fixing them is secondary to maintaining full compatibility with Word.

Security

If you find a security vulnerability, please do not open a public issue. Report it privately through GitHub Security Advisories.

License

This project is licensed under the MIT License - see the LICENSE file for details