diff --git a/.gitignore b/.gitignore
index 2f5e4a37b8a..e59c1b4349f 100644
--- a/.gitignore
+++ b/.gitignore
@@ -29,8 +29,8 @@ dlldata.c
 .sconsign.dblite
 user_docs/*/*.html
 user_docs/*/*.css
-user_docs/*/keyCommands.t2t
-user_docs/build.t2tConf
+# TODO remove when migrating from t2t to md as markdown files will be committed.
+user_docs/*/*.md
 extras/controllerClient/x86/nvdaController.h
 extras/controllerClient/x64
 extras/controllerClient/arm64
diff --git a/keyCommandsDoc.py b/keyCommandsDoc.py
index 170c3dfd72d..514788165eb 100644
--- a/keyCommandsDoc.py
+++ b/keyCommandsDoc.py
@@ -1,181 +1,186 @@
 # -*- coding: UTF-8 -*-
-#keyCommandsDoc.py
-#A part of NonVisual Desktop Access (NVDA)
-#This file is covered by the GNU General Public License.
-#See the file COPYING for more details.
-#Copyright (C) 2010-2019 NV Access Limited, Mesar Hameed, Takuya Nishimoto
+# A part of NonVisual Desktop Access (NVDA)
+# Copyright (C) 2023-2024 NV Access Limited, Mesar Hameed, Takuya Nishimoto
+# This file is covered by the GNU General Public License.
+# See the file COPYING for more details.
 
-"""Utilities related to NVDA Key Commands documents.
 """
+TODO: move to site_scons/site_tools
 
-import os
-import codecs
+Generates the Key Commands document from the User Guide.
+Works as a Python Markdown Extension:
+https://python-markdown.github.io/extensions/
+
+TODO move docs to docs folder
+
+Generation of the Key Commands document requires certain commands to be included in the user guide.
+These commands must begin at the start of the line and take the form::
+	%kc:command: arg
+
+The kc:title command must appear first and specifies the title of the Key Commands document.
+For example:
+	%kc:title: NVDA Key Commands
+
+The rest of these commands are used to include key commands into the document.
+Appropriate headings from the User Guide will be included implicitly.
+
+The kc:beginInclude command begins a block of text which should be included verbatim.
+The block ends at the kc:endInclude command.
+For example::
+	%kc:beginInclude
+	|| Name | Desktop command | Laptop command | Description |
+	...
+	%kc:endInclude
+
+The kc:settingsSection command indicates the beginning of a section documenting individual settings.
+It specifies the header row for a table summarising the settings indicated by the kc:setting command
+(see below).
+In order, it must consist of a name column, a column for each keyboard layout and a description column.
+For example::
+	%kc:settingsSection: || Name | Desktop command | Laptop command | Description |
+
+The kc:setting command indicates a section for an individual setting.
+It must be followed by:
+	* A heading containing the name of the setting;
+	* A table row for each keyboard layout, or if the key is common to all layouts,
+	a single line of text specifying the key after a colon;
+	* A blank line; and
+	* A line describing the setting.
+For example:
+	%kc:setting
+	==== Braille Tethered To ====
+	| Desktop command | NVDA+control+t |
+	| Laptop Command | NVDA+control+t |
+	This option allows you to choose whether the braille display will follow the system focus,
+	or whether it follows the navigator object / review cursor.
+"""
+
+from enum import auto, Enum, IntEnum, StrEnum
 import re
-import txt2tags
+from collections.abc import Iterator
+
+from markdown import Extension, Markdown
+from markdown.preprocessors import Preprocessor
+
+
+LINE_END = "\r\n"
+
+
+class Section(IntEnum):
+	"""Sections must be nested in this order."""
+	HEADER = auto()
+	BODY = auto()
+
+
+class Command(StrEnum):
+	TITLE = "title"
+	BEGIN_INCLUDE = "beginInclude"
+	END_INCLUDE = "endInclude"
+	SETTING = "setting"
+	SETTINGS_SECTION = "settingsSection"
+
+	def t2tRegex(self) -> re.Pattern:
+		return re.compile(rf"%kc:({self.value}.*)")
+
+
+class Regex(Enum):
+	COMMAND = re.compile(r"^<!-- KC:(?P<cmd>[^:\s]+)(?:: (?P<arg>.*))? -->$")
+	HEADING = re.compile(r"^(?P<id>#+)(?P<txt>.*)$")
+	SETTING_SINGLE_KEY = re.compile(r"^[^|]+?[:：]\s*(.+?)\s*$")
+	TABLE_ROW = re.compile(r"^(\|.*\|)$")
 
-LINE_END = u"\r\n"
 
 class KeyCommandsError(Exception):
 	"""Raised due to an error encountered in the User Guide related to generation of the Key Commands document.
 	"""
 
-class KeyCommandsMaker(object):
-	"""Generates the Key Commands document from the User Guide.
-	To generate a Key Commands document, create an instance and then call L{make} on it.
-	
-	Generation of the Key Commands document requires certain commands to be included in the user guide.
-	These commands must begin at the start of the line and take the form::
-		%kc:command: arg
-	
-	The kc:title command must appear first and specifies the title of the Key Commands document. For example::
-		%kc:title: NVDA Key Commands
-	
-	The kc:includeconf command allows you to insert a txt2tags includeconf command in the Key Commands document. For example::
-		%kc:includeconf: ../ar.t2tconf
-	You may use multiple kc:includeconf commands, but they must appear before any of the other commands below.
-	
-	The rest of these commands are used to include key commands into the document.
-	Appropriate headings from the User Guide will be included implicitly.
-	
-	The kc:beginInclude command begins a block of text which should be included verbatim.
-	The block ends at the kc:endInclude command.
-	For example::
-		%kc:beginInclude
-		|| Name | Desktop command | Laptop command | Description |
-		...
-		%kc:endInclude
-	
-	The kc:settingsSection command indicates the beginning of a section documenting individual settings.
-	It specifies the header row for a table summarising the settings indicated by the kc:setting command (see below).
-	In order, it must consist of a name column, a column for each keyboard layout and a description column.
-	For example::
-		%kc:settingsSection: || Name | Desktop command | Laptop command | Description |
-	
-	The kc:setting command indicates a section for an individual setting.
-	It must be followed by:
-		* A heading containing the name of the setting;
-		* A table row for each keyboard layout, or if the key is common to all layouts, a single line of text specifying the key after a colon;
-		* A blank line; and
-		* A line describing the setting.
-	For example::
-		%kc:setting
-		==== Braille Tethered To ====
-		| Desktop command | NVDA+control+t |
-		| Laptop Command | NVDA+control+t |
-		This option allows you to choose whether the braille display will follow the system focus, or whether it follows the navigator object / review cursor.
-	"""
 
-	t2tRe = None
-	RE_COMMAND = re.compile(r"^%kc:(?P<cmd>[^:\s]+)(?:: (?P<arg>.*))?$")
-	KCSECT_HEADER = 0
-	KCSECT_CONFIG = 1
-	KCSECT_BODY = 2
+class KeyCommandsExtension(Extension):
+	# Magic number, priorities are not well documented.
+	# It's unclear what the range of priorities are to compare to, but 25 seems to work, 1 doesn't.
+	# See https://python-markdown.github.io/extensions/api/#registries
+	PRIORITY = 25
 
-	@classmethod
-	def _initClass(cls):
-		if cls.t2tRe:
-			return
-		# Only fetch this once.
-		cls.t2tRe = txt2tags.getRegexes()
-
-	def __init__(self, userGuideFilename,keyCommandsFileName):
-		"""Constructor.
-		@param userGuideFilename: The file name of the User Guide to be used as input.
-		@type userGuideFilename: str
-		@param keyCommandsFilename: The file name of the key commands file to be output. 
-		@type keyCommandsFilename: str
-		"""
-		self._initClass()
-		self.ugFn = userGuideFilename
-		#: The file name of the Key Commands document that will be generated.
-		#: This will be in the same directory as the User Guide.
-		self.kcFn = keyCommandsFileName
+	def extendMarkdown(self, md: Markdown):
+		md.preprocessors.register(KeyCommandsPreprocessor(md), 'key_commands', self.PRIORITY)
+
+
+class KeyCommandsPreprocessor(Preprocessor):
+	def __init__(self, md: Markdown | None):
+		super().__init__(md)
+		self.initialize()
+
+	def initialize(self):
+		self._ugLines: Iterator[str] = iter(())
+		self._kcLines: list[str] = []
 		#: The current section of the key commands file.
-		self._kcSect = self.KCSECT_HEADER
+		self._kcSect: Section = Section.HEADER
 		#: The current stack of headings.
-		self._headings = []
+		self._headings: list[re.Match] = []
 		#: The 0 based level of the last heading in L{_headings} written to the key commands file.
-		self._kcLastHeadingLevel = -1
+		self._kcLastHeadingLevel: int = -1
 		#: Whether lines which aren't commands should be written to the key commands file as is.
-		self._kcInclude = False
+		self._kcInclude: bool = False
 		#: The header row for settings sections.
-		self._settingsHeaderRow = None
+		self._settingsHeaderRow: str | None = None
 		#: The number of layouts for settings in a settings section.
-		self._settingsNumLayouts = 0
+		self._settingsNumLayouts: int = 0
 		#: The current line number being processed, used to present location of syntax errors
-		self._lineNum = 0
-
-	def make(self):
-		"""Generate the Key Commands document.
-		@postcondition: If the User Guide contains appropriate commands, the Key Commands document will be generated and saved as L{kcFn}.
-			Otherwise, no file will be generated.
-		@return: C{True} if a document was generated, C{False} otherwise.
-		@rtype: bool
-		@raise IOError:
-		@raise KeyCommandsError:
-		"""
-		tKcFn=self.kcFn+'__'
-		self._ug = codecs.open(self.ugFn, "r", "utf-8-sig")
-		self._kc = codecs.open(tKcFn, "w", "utf-8-sig")
-
-		success=False
-		with self._ug, self._kc:
-			self._make()
-			success=self._kc.tell() > 0
-		if success:
-			os.rename(tKcFn,self.kcFn)
-		else:
-			os.remove(tKcFn)
-		return success
+		self._lineNum: int = 0
+		# We want to skip the title line to replace it with the KC:TITLE command argument.
+		self._skippedTitle = False
 
-	def _make(self):
-		for line in self._ug:
+	def run(self, lines: list[str]) -> list[str]:
+		# Turn this into an iterator so we can use next() to seek through lines.
+		self._ugLines = iter(lines)
+		for line in self._ugLines:
+			line = line.strip()
 			self._lineNum += 1
-			line = line.rstrip()
-			m = self.RE_COMMAND.match(line)
+
+			# We want to skip the title line to replace it with the KC:TITLE command argument.
+			if line.startswith("# ") and not self._skippedTitle:
+				self._skippedTitle = True
+				continue
+
+			m = Regex.COMMAND.value.match(line)
 			if m:
 				self._command(**m.groupdict())
 				continue
 
-			m = self.t2tRe["numtitle"].match(line)
+			m = Regex.HEADING.value.match(line)
 			if m:
 				self._heading(m)
 				continue
 
 			if self._kcInclude:
-				self._kc.write(line + LINE_END)
+				self._kcLines.append(line)
 
-	def _command(self, cmd=None, arg=None):
+		return self._kcLines.copy()
+
+	def _command(self, cmd: Command | None = None, arg: str | None = None):
 		# Handle header commands.
-		if cmd == "title":
-			if self._kcSect > self.KCSECT_HEADER:
-				raise KeyCommandsError("%d, title command is not valid here" % self._lineNum)
+		if cmd == Command.TITLE.value:
+			if self._kcSect > Section.HEADER:
+				raise KeyCommandsError(f"{self._lineNum}, title command is not valid here")
 			# Write the title and two blank lines to complete the txt2tags header section.
-			self._kc.write(arg + LINE_END * 3)
-			self._kcSect = self.KCSECT_CONFIG
-			self._kc.write("%%!includeconf: ../global.t2tconf%s" % LINE_END)
-			return
-		elif self._kcSect == self.KCSECT_HEADER:
-			raise KeyCommandsError("%d, title must be the first command" % self._lineNum)
-		elif cmd == "includeconf":
-			if self._kcSect > self.KCSECT_CONFIG:
-				raise KeyCommandsError("%d, includeconf command is not valid here" % self._lineNum)
-			self._kc.write("%%!includeconf: %s%s" % (arg, LINE_END))
+			self._kcLines.append("# " + arg + LINE_END * 2)
+			self._kcSect = Section.BODY
 			return
-		elif self._kcSect == self.KCSECT_CONFIG:
-			self._kc.write(LINE_END)
-			self._kcSect = self.KCSECT_BODY
 
-		if cmd == "beginInclude":
+		elif self._kcSect == Section.HEADER:
+			raise KeyCommandsError(f"{self._lineNum}, title must be the first command")
+
+		if cmd == Command.BEGIN_INCLUDE.value:
 			self._writeHeadings()
 			self._kcInclude = True
-		elif cmd == "endInclude":
+		elif cmd == Command.END_INCLUDE.value:
 			self._kcInclude = False
-			self._kc.write(LINE_END)
+			self._kcLines.append("")
 
-		elif cmd == "settingsSection":
+		elif cmd == Command.SETTINGS_SECTION.value:
 			# The argument is the table header row for the settings section.
-			self._settingsHeaderRow = arg
+			# Replace t2t header syntax with markdown syntax.
+			self._settingsHeaderRow = arg.replace("||", "|")
 			# There are name and description columns.
 			# Each of the remaining columns provides keystrokes for one layout.
 			# There's one less delimiter than there are columns, hence subtracting 1 instead of 2.
@@ -185,30 +190,37 @@ def _command(self, cmd=None, arg=None):
 					f"{self._lineNum}, settingsSection command must specify the header row for a table"
 					" summarising the settings"
 				)
-		elif cmd == "setting":
+
+		elif cmd == Command.SETTING.value:
 			self._handleSetting()
 
 		else:
-			raise KeyCommandsError("%d, Invalid command %s" % (self._lineNum, cmd))
+			raise KeyCommandsError(f"{self._lineNum}, Invalid command {cmd}")
+
+	def _seekNonEmptyLine(self) -> str:
+		"""Seeks to the next non-empty line in the user guide.
+		"""
+		line = next(self._ugLines).strip()
+		self._lineNum += 1
+		while not line:
+			try:
+				line = next(self._ugLines).strip()
+			except StopIteration:
+				return line
+			self._lineNum += 1
+		return line
 
-	def _areHeadingsPending(self):
+	def _areHeadingsPending(self) -> bool:
 		return self._kcLastHeadingLevel < len(self._headings) - 1
 
 	def _writeHeadings(self):
 		level = self._kcLastHeadingLevel + 1
 		# Only write headings we haven't yet written.
 		for level, heading in enumerate(self._headings[level:], level):
-			# We don't want numbered headings in the output.
-			label=heading.group("label")
-			headingText = u"{id}{txt}{id}{label}".format(
-				id="=" * len(heading.group("id")),
-				txt=heading.group("txt"),
-				label="[%s]" % label if label else "")
-			# Write the heading and a blank line.
-			self._kc.write(headingText + LINE_END * 2)
+			self._kcLines.append(heading.group(0))
 		self._kcLastHeadingLevel = level
 
-	def _heading(self, m):
+	def _heading(self, m: re.Match):
 		# We work with 0 based heading levels.
 		level = len(m.group("id")) - 1
 		try:
@@ -218,7 +230,6 @@ def _heading(self, m):
 		self._headings.append(m)
 		self._kcLastHeadingLevel = min(self._kcLastHeadingLevel, level - 1)
 
-	RE_SETTING_SINGLE_KEY = re.compile(r"^[^|]+?[:：]\s*(.+?)\s*$")
 	def _handleSetting(self):
 		if not self._settingsHeaderRow:
 			raise KeyCommandsError("%d, setting command cannot be used before settingsSection command" % self._lineNum)
@@ -226,60 +237,65 @@ def _handleSetting(self):
 		if self._areHeadingsPending():
 			# There are new headings to write.
 			# If there was a previous settings table, it ends here, so write a blank line.
-			self._kc.write(LINE_END)
+			self._kcLines.append("")
 			self._writeHeadings()
 			# New headings were written, so we need to output the header row.
-			self._kc.write(self._settingsHeaderRow + LINE_END)
+			self._kcLines.append(self._settingsHeaderRow)
+			numCols = self._settingsNumLayouts + 2  # name + description + layouts
+			self._kcLines.append("|" + "---|" * numCols)
 
 		# The next line should be a heading which is the name of the setting.
-		line = next(self._ug)
-		self._lineNum += 1
-		m = self.t2tRe["title"].match(line)
+		line = self._seekNonEmptyLine()
+		m = Regex.HEADING.value.match(line)
 		if not m:
-			raise KeyCommandsError("%d, setting command must be followed by heading" % self._lineNum)
+			raise KeyCommandsError(f"{self._lineNum}, setting command must be followed by heading")
 		name = m.group("txt")
 
 		# The next few lines should be table rows for each layout.
 		# Alternatively, if the key is common to all layouts, there will be a single line of text specifying the key after a colon.
-		keys = []
-		for layout in range(self._settingsNumLayouts):
-			line = next(self._ug).strip()
-			self._lineNum += 1
-			m = self.RE_SETTING_SINGLE_KEY.match(line)
+		keys: list[str] = []
+		for _layout in range(self._settingsNumLayouts):
+			line = self._seekNonEmptyLine()
+
+			m = Regex.SETTING_SINGLE_KEY.value.match(line)
 			if m:
 				keys.append(m.group(1))
 				break
-			elif not self.t2tRe["table"].match(line):
-				raise KeyCommandsError("%d, setting command: There must be one table row for each keyboard layout" % self._lineNum)
+			elif not Regex.TABLE_ROW.value.match(line):
+				raise KeyCommandsError(
+					f"{self._lineNum}, setting command: "
+					"There must be one table row for each keyboard layout"
+				)
+
 			# This is a table row.
 			# The key will be the second column.
-			# TODO: Error checking.
-			key = line.strip("|").split("|")[1].strip()
-			keys.append(key)
+			try:
+				key = line.strip("|").split("|")[1].strip()
+			except IndexError:
+				raise KeyCommandsError(f"{self._lineNum}, setting command: Key entry not found in table row.")
+			else:
+				keys.append(key)
+
 		if 1 == len(keys) < self._settingsNumLayouts:
 			# The key has only been specified once, so it is the same in all layouts.
 			key = keys[0]
-			keys[1:] = (key for layout in range(self._settingsNumLayouts - 1))
+			keys[1:] = (key for _layout in range(self._settingsNumLayouts - 1))
 
 		# There should now be a blank line.
-		line = next(self._ug).strip()
+		line = next(self._ugLines).strip()
 		self._lineNum += 1
 		if line:
-			raise KeyCommandsError("%d, setting command: The keyboard shortcuts must be followed by a blank line. Multiple keys must be included in a table. Erroneous key: %s" % (self._lineNum, key))
+			raise KeyCommandsError(
+				f"{self._lineNum}, setting command: The keyboard shortcuts must be followed by a blank line. "
+				"Multiple keys must be included in a table. "
+				f"Erroneous key: {key}"
+			)
 
 		# Finally, the next line should be the description.
-		desc = next(self._ug).strip()
-		self._lineNum += 1
-
-		self._kc.write(u"| {name} | {keys} | {desc} |{lineEnd}".format(
-			name=name,
-			keys=u" | ".join(keys),
-			desc=desc, lineEnd=LINE_END))
-
-	def remove(self):
-		"""Remove the generated Key Commands document.
-		"""
-		try:
-			os.remove(self.kcFn)
-		except OSError:
-			pass
+		desc = self._seekNonEmptyLine()
+		self._kcLines.append(f"| {name} | {' | '.join(keys)} | {desc} |")
+		if not self._kcLines[-2].startswith("|"):
+			# The previous line was not a table, so this is a new table.
+			# Write the header row.
+			numCols = len(keys) + 2  # name + description + layouts
+			self._kcLines.append("|" + "---|" * numCols)
diff --git a/miscDeps b/miscDeps
index 0858efa727e..ec251b62be4 160000
--- a/miscDeps
+++ b/miscDeps
@@ -1 +1 @@
-Subproject commit 0858efa727e8eedc7e0ce4c159ae6a232234cf10
+Subproject commit ec251b62be456860ddb3b9c0346b1ea0dcc7fd67
diff --git a/projectDocs/dev/developerGuide/developerGuide.t2t b/projectDocs/dev/developerGuide/developerGuide.t2t
index 2fbd33829cd..0b9d7ced24f 100644
--- a/projectDocs/dev/developerGuide/developerGuide.t2t
+++ b/projectDocs/dev/developerGuide/developerGuide.t2t
@@ -3,9 +3,6 @@ NVDA NVDA_VERSION Developer Guide
 
 %!includeconf: ../../../user_docs/userGuide.t2tconf
 
-% Remove double spacing from the beginning of each line as txt2tags seems to indent preformatted text by two spaces
-%!PostProc(html): '^  ' ''
-
 = Table of Contents =[toc]
 %%toc
 
diff --git a/projectDocs/dev/developerGuide/sconscript b/projectDocs/dev/developerGuide/sconscript
index 46776b3bec9..c1dba7c73be 100644
--- a/projectDocs/dev/developerGuide/sconscript
+++ b/projectDocs/dev/developerGuide/sconscript
@@ -5,28 +5,36 @@
 
 import sys
 
-Import("env", "outputDir", "sourceDir", "t2tBuildConf")
+Import("env", "outputDir", "sourceDir")
 
 env = env.Clone()
 
 devDocsOutputDir=outputDir.Dir('devDocs')
 
 #Build the developer guide and move it to the output directory
-htmlFile=env.txt2tags('developerGuide.t2t')
-env.Depends(htmlFile, t2tBuildConf)
-env.SideEffect('_txt2tags',htmlFile)
+mdFile = env.txt2tags('developerGuide.t2t')
+env.SideEffect('_txt2tags', mdFile)
+htmlFile = env.md2html('developerGuide.md')
+env.Depends(htmlFile, mdFile)
+devGuideMD = env.Command(
+	target=devDocsOutputDir.File('developerGuide.md'),
+	source=mdFile,
+	action=Move('$TARGET', '$SOURCE')
+)
+env.Depends(devGuideMD, mdFile)
 devGuide = env.Command(
 	target=devDocsOutputDir.File('developerGuide.html'),
 	source=htmlFile,
-	action=Move('$TARGET','$SOURCE')
+	action=Move('$TARGET', '$SOURCE')
 )
+env.Depends(devGuide, devGuideMD)
 env.Alias("developerGuide",devGuide)
 
 devDocs_nvdaHelper_temp=env.Doxygen(source='../../../nvdaHelper/doxyfile')
 devDocs_nvdaHelper = env.Command(
 	target=devDocsOutputDir.Dir('nvdaHelper'),
 	source=devDocs_nvdaHelper_temp,
-	action=Move('$TARGET','$SOURCE')
+	action=Move('$TARGET', '$SOURCE')
 )
 env.Alias('devDocs_nvdaHelper', devDocs_nvdaHelper)
 env.Clean('devDocs_nvdaHelper', devDocs_nvdaHelper)
diff --git a/requirements.txt b/requirements.txt
index a27a4f916b3..962e26bd7b7 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -19,6 +19,12 @@ git+https://github.com/py2exe/py2exe@4e7b2b2c60face592e67cb1bc935172a20fa371d#eg
 # Creating XML unit test reports
 unittest-xml-reporting==3.2.0
 
+# Building user documentation
+Markdown==3.5.1
+mdx_truly_sane_lists==1.3
+markdown-link-attr-modifier==0.2.1
+mdx-gh-links==0.4
+
 # For building developer documentation
 sphinx==7.2.6
 sphinx_rtd_theme==1.3.0
diff --git a/sconstruct b/sconstruct
index d5af27f7485..0aefe68a08d 100755
--- a/sconstruct
+++ b/sconstruct
@@ -54,7 +54,6 @@ sys.path.append(sourceEnvPath)
 import sourceEnv
 sys.path.remove(sourceEnvPath)
 import time
-from glob import glob
 import importlib.util
 import winreg
 
@@ -84,20 +83,7 @@ makensis = os.path.abspath(os.path.join("include", "nsis", "NSIS", "makensis.exe
 # Get the path to xgettext.
 XGETTEXT = os.path.abspath(os.path.join("miscDeps", "tools", "xgettext.exe"))
 
-def keyCommandsDocTool(env):
-	import keyCommandsDoc
-	kcdAction=env.Action(
-		lambda target,source,env: not keyCommandsDoc.KeyCommandsMaker(source[0].path,target[0].path).make(),
-		lambda target,source,env: 'Generating %s'%target[0],
-	)
-	kcdBuilder=env.Builder(
-		action=kcdAction,
-		suffix='.t2t',
-		src_suffix='.t2t',
-	)
-	env['BUILDERS']['keyCommandsDoc']=kcdBuilder
-
-keyCommandsLangBlacklist=set([])
+import txt2tags  # noqa: F401, import required here to build docs
 
 vars = Variables()
 vars.Add("version", "The version of this build", versionInfo.version)
@@ -117,8 +103,8 @@ vars.Add(EnumVariable('nvdaHelperLogLevel','The level of logging you wish to see
 env = Environment(variables=vars,HOST_ARCH='x86',tools=[
 	"textfile",
 	"gettextTool",
-	"t2t"
-	,keyCommandsDocTool,
+	"t2t",
+	"md2html",
 	"doxygen",
 	"recursiveInstall"
 ])
@@ -248,19 +234,6 @@ envArm64.SConscript('nvdaHelper/archBuild_sconscript',exports={'env':envArm64,'c
 for po in env.Glob(sourceDir.path+'/locale/*/lc_messages/*.po'):
 	env.gettextMoFile(po)
 
-#Allow all key command t2t files to be generated from their userGuide t2t sources
-for lang in os.listdir(userDocsDir.path):
-	if lang in keyCommandsLangBlacklist: continue
-	for ug in glob(os.path.join(userDocsDir.path,lang,'userGuide.t2t')):
-		env.keyCommandsDoc(File(ug).File('keyCommands.t2t'),ug)
-
-t2tBuildConf = env.Textfile(userDocsDir.File("build.t2tConf"), [
-	# We need to do this one as PostProc so it gets converted for the title.
-	r"%!PostProc: NVDA_VERSION {}".format(version),
-	r"%!PreProc: NVDA_URL {}".format(versionInfo.url),
-	r"%!PreProc: NVDA_COPYRIGHT_YEARS {}".format(versionInfo.copyrightYears),
-])
-
 userGuideConf = os.path.join(userDocsDir.path, 'userGuide.t2tconf')
 globalConf = os.path.join(userDocsDir.path, 'global.t2tconf')
 changesConf = os.path.join(userDocsDir.path, 'changes.t2tconf')
@@ -269,22 +242,33 @@ styles = os.path.join(userDocsDir.path, 'styles.css')
 #Allow all t2t files to be converted to html in user_docs
 #As we use scons Glob this will also include the keyCommands.t2t files
 for t2tFile in env.Glob(os.path.join(userDocsDir.path,'*','*.t2t')):
-	htmlFile=env.txt2tags(t2tFile)
+	mdFile = env.txt2tags(t2tFile)
 	#txt2tags can not be run in parallel so make sure scons knows this
-	env.SideEffect('_txt2tags',htmlFile)
-	# All of our t2t files require build.t2tConf.
-	styleInstallPath = os.path.dirname(t2tFile.abspath)
-	installedStyle = env.Install(styleInstallPath, styles)
+	env.SideEffect('_txt2tags', mdFile)
 	env.Depends(
-		htmlFile,
+		mdFile,
 		[
-			t2tBuildConf,
 			userGuideConf,
 			globalConf,
 			changesConf,
+		]
+	)
+	htmlFile = env.md2html(t2tFile.abspath.replace(".t2t", ".md"))
+	styleInstallPath = os.path.dirname(t2tFile.abspath)
+	installedStyle = env.Install(styleInstallPath, styles)
+	env.Depends(
+		htmlFile,
+		[
 			styles,
-			installedStyle
-		])
+			installedStyle,
+		]
+	)
+	env.Depends(htmlFile, mdFile)
+
+# Create key commands files
+for userGuideFile in env.Glob(os.path.join(userDocsDir.path, '*', 'userGuide.md')):
+	keyCommandsHtmlFile = env.md2html(userGuideFile.abspath.replace("userGuide.md", "keyCommands.html"), userGuideFile)
+	env.Depends(keyCommandsHtmlFile, userGuideFile)
 
 # Build unicode CLDR dictionaries
 env.SConscript('cldrDict_sconscript',exports=['env', 'sourceDir'])
@@ -430,12 +414,24 @@ env.Alias("client", clientArchive)
 
 outputStylesFile=env.Command(outputDir.File("styles.css"),userDocsDir.File('styles.css'),Copy('$TARGET','$SOURCE'))
 changesFile=env.Command(outputDir.File("%s_changes.html" % outFilePrefix),userDocsDir.File('en/changes.html'),Copy('$TARGET','$SOURCE'))
+changesMDFile = env.Command(
+	outputDir.File("changes.md"),
+	userDocsDir.File('en/changes.md'),
+	Copy('$TARGET','$SOURCE')
+)
+env.Depends(changesFile, changesMDFile)
 env.Depends(changesFile, outputStylesFile)
 env.Alias('changes',changesFile)
 
 userGuideFile=env.Command(outputDir.File("userGuide.html"),userDocsDir.File('en/userGuide.html'),Copy('$TARGET','$SOURCE'))
+userGuideMDFile = env.Command(
+	outputDir.File("userGuide.md"),
+	userDocsDir.File('en/userGuide.md'),
+	Copy('$TARGET', '$SOURCE')
+)
+env.Depends(userGuideFile, userGuideMDFile)
 env.Depends(userGuideFile, outputStylesFile)
-env.Alias('userGuide',userGuideFile)
+env.Alias('userGuide', userGuideFile)
 
 keyCommandsFile = env.Command(
 	outputDir.File("keyCommands.html"),
@@ -494,7 +490,7 @@ def makePot(target, source, env):
 	os.rename(tmpFn, potFn)
 
 
-env.SConscript("projectDocs/dev/developerGuide/sconscript", exports=["env", "outputDir", "sourceDir", "t2tBuildConf"])
+env.SConscript("projectDocs/dev/developerGuide/sconscript", exports=["env", "outputDir", "sourceDir"])
 
 
 def getSubDirs(path):
diff --git a/site_scons/site_tools/md2html.py b/site_scons/site_tools/md2html.py
new file mode 100644
index 00000000000..9048f760fc8
--- /dev/null
+++ b/site_scons/site_tools/md2html.py
@@ -0,0 +1,161 @@
+# A part of NonVisual Desktop Access (NVDA)
+# Copyright (C) 2023 NV Access Limited
+# This file may be used under the terms of the GNU General Public License, version 2 or later.
+# For more details see: https://www.gnu.org/licenses/gpl-2.0.html
+
+from importlib.util import find_spec
+import io
+import pathlib
+import re
+import shutil
+
+import SCons.Node.FS
+import SCons.Environment
+
+DEFAULT_EXTENSIONS = frozenset({
+	# Supports tables, HTML mixed with markdown, code blocks and more
+	"markdown.extensions.extra",
+	# Allows TOC with [TOC], allows anchors set by "# Title {#foo}"
+	"markdown.extensions.toc",
+	# Allows setting attributes with {@id=foo}
+	"markdown.extensions.legacy_attrs",
+	# Adds code highlighting to code blocks
+	"markdown.extensions.codehilite",
+	# Makes list behaviour better, including 2 space indents by default
+	"mdx_truly_sane_lists",
+	# External links will open in a new tab, and title will be set to the link text
+	"markdown_link_attr_modifier",
+	# Adds links to GitHub authors, issues and PRs
+	"mdx_gh_links",
+})
+
+_extensionConfigs = {
+	"markdown_link_attr_modifier": {
+		"new_tab": "external_only",
+		"auto_title": "on",
+	},
+	"mdx_gh_links": {
+		"user": "nvaccess",
+		"repo": "nvda",
+	},
+}
+
+_RTLlangCodes = {"ar", "fa", "he"}
+
+
+def _replaceNVDATags(md: str, env: SCons.Environment.Environment) -> str:
+	import versionInfo
+	# Replace tags in source file
+	md = md.replace("NVDA_VERSION", env["version"])
+	md = md.replace("NVDA_URL", versionInfo.url)
+	md = md.replace("NVDA_COPYRIGHT_YEARS", versionInfo.copyrightYears)
+	return md
+
+
+def _getTitle(mdBuffer: io.BytesIO, isKeyCommands: bool = False) -> str:
+	if isKeyCommands:
+		TITLE_RE = re.compile(r"^<!-- KC:title: (.*) -->$")
+		# Make next read at start of buffer
+		mdBuffer.seek(0)
+		for line in mdBuffer.readlines():
+			match = TITLE_RE.match(line.decode("utf-8").strip())
+			if match:
+				return match.group(1)
+
+		raise ValueError("No KC:title command found in userGuide.md")
+
+	else:
+		# Make next read at start of buffer
+		mdBuffer.seek(0)
+		# Remove heading hashes and trailing whitespace to get the tab title
+		title = mdBuffer.readline().decode("utf-8").strip().lstrip("# ")
+
+	return title
+
+
+def md2html_actionFunc(
+		target: list[SCons.Node.FS.File],
+		source: list[SCons.Node.FS.File],
+		env: SCons.Environment.Environment
+):
+	import markdown
+	isKeyCommands = target[0].path.endswith("keyCommands.html")
+
+	with open(source[0].path, "r", encoding="utf-8") as mdFile:
+		mdStr = mdFile.read()
+
+	mdStr = _replaceNVDATags(mdStr, env)
+
+	# Write replaced source to buffer.
+	# md has a bug with StringIO, so BytesIO is required.
+	mdBuffer = io.BytesIO()
+	mdBuffer.write(mdStr.encode("utf-8"))
+
+	lang = pathlib.Path(source[0].path).parent.name
+	title = _getTitle(mdBuffer, isKeyCommands)
+	# md has a bug with StringIO, so BytesIO is required.
+	htmlBuffer = io.BytesIO()
+	htmlBuffer.write(
+		f"""
+<!DOCTYPE html>
+<html lang="{lang}" dir="{"rtl" if lang in _RTLlangCodes else "ltr"}">
+<head>
+<meta charset="utf-8">
+<title>{title}</title>
+<meta name="viewport" content="width=device-width, initial-scale=1" />
+<link rel="stylesheet" href="styles.css">
+</head>
+<body>
+	""".strip().encode("utf-8")
+	)
+
+	# Make next read from start of buffer
+	mdBuffer.seek(0)
+	# Make next write append at end of buffer
+	htmlBuffer.seek(0, io.SEEK_END)
+
+	extensions = set(DEFAULT_EXTENSIONS)
+	if isKeyCommands:
+		from keyCommandsDoc import KeyCommandsExtension
+		extensions.add(KeyCommandsExtension())
+
+	markdown.markdownFromFile(
+		input=mdBuffer,
+		output=htmlBuffer,
+		# https://python-markdown.github.io/extensions/
+		extensions=extensions,
+		extension_configs=_extensionConfigs,
+	)
+
+	# Make next write append at end of buffer
+	htmlBuffer.seek(0, io.SEEK_END)
+	htmlBuffer.write(b"\n</body>\n</html>\n")
+
+	with open(target[0].path, "wb") as targetFile:
+		# Make next read at start of buffer
+		htmlBuffer.seek(0)
+		shutil.copyfileobj(htmlBuffer, targetFile)
+
+	mdBuffer.close()
+	htmlBuffer.close()
+
+
+def exists(env: SCons.Environment.Environment) -> bool:
+	for ext in [
+		"markdown",
+		"markdown_link_attr_modifier",
+		"mdx_truly_sane_lists",
+		"mdx_gh_links",
+		"keyCommandsDoc",
+	]:
+		if find_spec(ext) is None:
+			return False
+	return True
+
+
+def generate(env: SCons.Environment.Environment):
+	env["BUILDERS"]["md2html"] = env.Builder(
+		action=env.Action(md2html_actionFunc, lambda t, s, e: f"Converting {s[0].path} to {t[0].path}"),
+		suffix=".html",
+		src_suffix=".md"
+	)
diff --git a/site_scons/site_tools/t2t.py b/site_scons/site_tools/t2t.py
index 8f600723822..2eab5a34f30 100644
--- a/site_scons/site_tools/t2t.py
+++ b/site_scons/site_tools/t2t.py
@@ -1,31 +1,47 @@
-###
-#This file is a part of the NVDA project.
-#URL: http://www.nvda-project.org/
-#Copyright 2010 James Teh <jamie@jantrid.net>.
-#This program is free software: you can redistribute it and/or modify
-#it under the terms of the GNU General Public License version 2.0, as published by
-#the Free Software Foundation.
-#This program is distributed in the hope that it will be useful,
-#but WITHOUT ANY WARRANTY; without even the implied warranty of
-#MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
-#This license can be found at:
-#http://www.gnu.org/licenses/old-licenses/gpl-2.0.html
-###
-
-def txt2tags_actionFunc(target,source,env):
+# A part of NonVisual Desktop Access (NVDA)
+# Copyright (C) 2010-2024 NV Access Limited, James Teh
+# This file is covered by the GNU General Public License.
+# See the file COPYING for more details.
+
+import SCons
+
+
+def txt2tags_actionFunc(
+		target: list[SCons.Node.FS.File],
+		source: list[SCons.Node.FS.File],
+		env: SCons.Environment.Environment
+):
 	import txt2tags
-	txt2tags.exec_command_line([str(source[0])])
+	from keyCommandsDoc import Command
+
+	with open(source[0].path, "r", encoding="utf-8") as mdFile:
+		mdOriginal = mdFile.read()
+		mdStr = str(mdOriginal)
+
+	with open(source[0].path, "w", encoding="utf-8") as mdFile:
+		# Substitute t2t key commands with markdown comments temporarily
+		for command in Command:
+			mdStr = command.t2tRegex().sub(lambda m: f"<!-- KC:{m.group(1)} -->", mdStr)
+		mdFile.write(mdStr)
 
-def exists(env):
+	txt2tags.exec_command_line([source[0].path])
+
+	with open(source[0].path, "w", encoding="utf-8") as mdFile:
+		# Restore to original
+		mdFile.write(mdOriginal)
+
+
+def exists(env: SCons.Environment.Environment) -> bool:
 	try:
 		import txt2tags
 		return True
 	except ImportError:
 		return False
 
-def generate(env):
-	env['BUILDERS']['txt2tags']=env.Builder(
-		action=env.Action(txt2tags_actionFunc,lambda t,s,e: 'Converting %s to html'%s[0].path),
-		suffix='.html',
+
+def generate(env: SCons.Environment.Environment):
+	env["BUILDERS"]["txt2tags"] = env.Builder(
+		action=env.Action(txt2tags_actionFunc, lambda t, s, e: f"Converting {s[0].path} to md"),
+		suffix='.md',
 		src_suffix='.t2t'
 	)
diff --git a/user_docs/ar/locale.t2tconf b/user_docs/ar/locale.t2tconf
index 73483d6fb5c..e69de29bb2d 100644
--- a/user_docs/ar/locale.t2tconf
+++ b/user_docs/ar/locale.t2tconf
@@ -1,36 +0,0 @@
-%!PostProc(html): ^<HTML>$ <HTML LANG='AR' DIR='RTL'>
-%!PostProc(html): TEXT="black">$  TEXT='black' STYLE='font-family: Simplified Arabic'>
-%!PostProc(html): NVDA_AR NVDA
-
-% add rtl macro
-%!PostProc: %ltr{(.*)} <span dir='ltr'>\1</span>
-
-% reverse the numbering, to be displayed correctly in arabic.
-%PostProc(html): >([0-9]+)\.([0-9]+)\.([0-9]+)\.  > ar.\3ar.\2ar.\1
-%PostProc(html): >([0-9]+)\.([0-9]+)\. > ar.\2ar.\1
-%PostProc(html): >([0-9]+)\. > ar.\1
-
-% now replace Arabic numerals with indian once.
-% 10 11, ... etc should go before signle digits, otherwise they will interfear
-% and not produce the correct output.
-%PostProc(html): ar\.10 ١٠.
-%PostProc(html): ar\.11 ١١.
-%PostProc(html): ar\.12 ١٢.
-%PostProc(html): ar\.13 ١٣.
-%PostProc(html): ar\.14 ١٤.
-%PostProc(html): ar\.15 ١٥.
-%PostProc(html): ar\.16 ١٦.
-%PostProc(html): ar\.17 ١٧.
-%PostProc(html): ar\.18 ١٨.
-%PostProc(html): ar\.19 ١٩.
-
-%PostProc(html): ar\.0 ٠.
-%PostProc(html): ar\.1 ١.
-%PostProc(html): ar\2. ٢.
-%PostProc(html): ar\.3 .3
-%PostProc(html): ar\.4 ٤.
-%PostProc(html): ar\.5 ٥.
-%PostProc(html): ar\.6 ٦.
-%PostProc(html): ar\.7 ٧.
-%PostProc(html): ar\.8 ٨.
-%PostProc(html): ar\.9 ٩.
diff --git a/user_docs/ca/locale.t2tconf b/user_docs/ca/locale.t2tconf
index a1cb4bb37f2..e7144385b66 100644
--- a/user_docs/ca/locale.t2tconf
+++ b/user_docs/ca/locale.t2tconf
@@ -1,8 +1,5 @@
 % locale.t2tconf for Catalan
 
-% Set document language
-%!PostProc(html): '^<HTML' '<HTML LANG="ca"'
-
 % macro to allow hreflang attribute
 % rewrite en_[string url] to <a hreflang="en" href="url">string</a>
-%!postproc(html): '(?i)([a-z-_]+)_<a href' '<A HREFLANG="\1" HREF'
+%!PreProc: '([A-za-z-_]+)_\[([^\]]*)\]' '[{@hreflang=\1}\2]'
diff --git a/user_docs/changes.t2tconf b/user_docs/changes.t2tconf
index 2774198ec67..28bd981fd74 100644
--- a/user_docs/changes.t2tconf
+++ b/user_docs/changes.t2tconf
@@ -1,12 +1 @@
 %!includeconf: global.t2tconf
-
-% Make ticket references into links.
-%!PreProc: "#([0-9]+)" "[#\1 https://github.com/nvaccess/nvda/issues/\1]"
-
-% Make GitHub username references into links.
-% Github username may only contain alphanumeric characters or hyphens.
-% Maximum is 39 characters.
-%!PreProc: ", @([A-Za-z0-9\-]{1,39})" ", [@\1 https://github.com/nvaccess/nvda/commits?author=\1]"
-
-% Make links open in a new tab/window.
-%!PostProc(html): '(<A HREF=".*?")>' '\1 target="_blank">'
diff --git a/user_docs/de/developerGuide.t2t b/user_docs/de/developerGuide.t2t
index 1059be63fa9..02e78586f9d 100644
--- a/user_docs/de/developerGuide.t2t
+++ b/user_docs/de/developerGuide.t2t
@@ -3,9 +3,6 @@
 
 %!includeconf: ../userGuide.t2tconf
 
-% Remove double spacing from the beginning of each line as txt2tags seems to indent preformatted text by two spaces
-%!PostProc(html): '^  ' ''
-
 = Inhaltsverzeichnis =[toc]
 %%toc
 
diff --git a/user_docs/fa/locale.t2tconf b/user_docs/fa/locale.t2tconf
index 92284f575de..e69de29bb2d 100644
--- a/user_docs/fa/locale.t2tconf
+++ b/user_docs/fa/locale.t2tconf
@@ -1,37 +0,0 @@
-%!PostProc(html): ^<HTML>$ <HTML LANG='FA' DIR='RTL'>
-%!PostProc(html): TEXT="black">$  TEXT='black' STYLE='font-family: Simplified Arabic'>
-
-% add rtl macro
-%!PostProc: %ltr{(.*)} <span dir='ltr'>\1</span>
-
-% reverse the numbering, to be displayed correctly in Persian.
-%!PostProc(html): >([0-9]+)\.([0-9]+)\.([0-9]+)\.  > fa.\3fa.\2fa.\1
-%!PostProc(html): >([0-9]+)\.([0-9]+)\. > fa.\2fa.\1
-%!PostProc(html): >([0-9]+)\. > fa.\1
-
-% now replace Persian numerals with indian ones.
-% 10 11, ... etc should go before signle digits, otherwise they will interfear
-% and not produce the correct output.
-%!PostProc(html): fa\.10 ۱۰.
-%!PostProc(html): fa\.11 ۱۱.
-%!PostProc(html): fa\.12 ۱۲.
-%!PostProc(html): fa\.13 ۱۳.
-%!PostProc(html): fa\.14 ۱۴.
-%!PostProc(html): fa\.15 ۱۵.
-%!PostProc(html): fa\.16 ۱۶.
-%!PostProc(html): fa\.17 ۱۷.
-%!PostProc(html): fa\.18 ۱۸.
-%!PostProc(html): fa\.19 ۱۹.
-%!PostProc(html): fa\.20 ۲۰.
-%!PostProc(html): fa\.21 ۲۱.
-
-%!PostProc(html): fa\.0 ۰.
-%!PostProc(html): fa\.1 ۱.
-%!PostProc(html): fa\.2 ۲.
-%!PostProc(html): fa\.3 ۳.
-%!PostProc(html): fa\.4 ۴.
-%!PostProc(html): fa\.5 ۵.
-%!PostProc(html): fa\.6 ۶.
-%!PostProc(html): fa\.7 ۷.
-%!PostProc(html): fa\.8 ۸.
-%!PostProc(html): fa\.9 ۹.
diff --git a/user_docs/global.t2tconf b/user_docs/global.t2tconf
index 5757bfe8b8d..7601f7360b0 100644
--- a/user_docs/global.t2tconf
+++ b/user_docs/global.t2tconf
@@ -1,24 +1,6 @@
-%!Target: html
+%!Target: md
 %!Encoding: UTF-8
 
 % Remove the Table of Contents heading from the toc.
-%!PostProc(html): '^.*\<LI\>\<A HREF="#toc"\>.*\</A\>.*$' ''
-
-% h1 in html should really be the document title only.
-% Therefore, change h1 through h5 in the output to h2 through h6.
-%!PostProc(html): ^<H5>(.*)</H5>$ <H6>\1</H6>
-%!PostProc(html): ^<H4>(.*)</H4>$ <H5>\1</H5>
-%!PostProc(html): ^<H3>(.*)</H3>$ <H4>\1</H4>
-%!PostProc(html): ^<H2>(.*)</H2>$ <H3>\1</H3>
-%!PostProc(html): ^<H1>(.*)</H1>$ <H2>\1</H2>
-
-% Some of our files contain the UTF-8 BOM.
-% txt2tags doesn't care about encodings internally,
-% so it will just include the BOM at the start of the title.
-% Therefore, strip the BOM from the title.
-%!PostProc(html): \<(TITLE|H1)\>\xef\xbb\xbf <\1>
-
-%!Options: --style styles.css
-
-% This provides the macros NVDA_VERSION, NVDA_URL and NVDA_COPYRIGHT_YEARS.
-%!includeconf: build.t2tconf
+% For some reason t2t to md conversion always forces the TOC at the top of the document 
+%!PostProc(md): '^#.*\{#toc\}.*$' ''
diff --git a/user_docs/pt_BR/developerGuide.t2t b/user_docs/pt_BR/developerGuide.t2t
index 040a388f791..4eb6d7f77b0 100644
--- a/user_docs/pt_BR/developerGuide.t2t
+++ b/user_docs/pt_BR/developerGuide.t2t
@@ -3,9 +3,6 @@
 
 %!includeconf: ../userGuide.t2tconf
 
-% Remove double spacing from the beginning of each line as txt2tags seems to indent preformatted text by two spaces
-%!PostProc(html): '^  ' ''
-
 Nota; Alguns termos deverão ser mantidos em Inglês por referirem-se a programação e Jargões Técnicos
 
 
diff --git a/user_docs/styles.css b/user_docs/styles.css
index 76311cfe121..f8e25ac6072 100644
--- a/user_docs/styles.css
+++ b/user_docs/styles.css
@@ -7,6 +7,10 @@ body {
   color: #333;
 }
 
+:lang(ar) body *, :lang(fa) body * {
+  font-family: "Simplified Arabic", "Traditional Arabic", "Arial", sans-serif !important;
+}
+
 h1 {
   color:white;
   background-color: #472F5F; /* NVDA purple */
diff --git a/user_docs/userGuide.t2tconf b/user_docs/userGuide.t2tconf
index 8885ab8e723..a5b5f11d7e6 100644
--- a/user_docs/userGuide.t2tconf
+++ b/user_docs/userGuide.t2tconf
@@ -1,3 +1,3 @@
 %!includeconf: global.t2tconf
 
-%!Options: --toc
+%!Options: --toc --no-enum-title