diff --git a/README.md b/README.md
index 4b24dbd..a648218 100644
--- a/README.md
+++ b/README.md
@@ -12,6 +12,7 @@ CLI tool to query LDAP/AD servers
   * JSON stream (with detailed or simplified attribute values)
   * CSV
   * Markdown table with stretched columns (for viewing in CLI/for monospaces fonts)
+  * HTML
 * Decodes certain well-known attributes (UUIDs, Timestamps, SID, userAccountControl)
 * Requires server to support [RFC 2696: Simple Paged Results](https://www.rfc-editor.org/rfc/rfc2696) for proper pagination
   * By default the first 1000 entries are shown, and it errors if there are more results
diff --git a/src/ldaptool/_main.py b/src/ldaptool/_main.py
index da67c45..60f081e 100644
--- a/src/ldaptool/_main.py
+++ b/src/ldaptool/_main.py
@@ -4,6 +4,7 @@ import argparse
 import csv
 import dataclasses
 import enum
+import html
 import subprocess
 import sys
 import typing
@@ -16,6 +17,13 @@ from ldaptool._utils.ldap import Result, SizeLimitExceeded
 class TableOutput(enum.StrEnum):
     MARKDOWN = "markdown"
     CSV = "csv"
+    HTML = "html"
+
+
+def _html_escape_line(columns: typing.Sequence[str], *, cell: str = "td") -> str:
+    cell_s = f"<{cell}>"
+    cell_e = f"</{cell}>"
+    return "<tr>" + ("".join(cell_s + html.escape(col) + cell_e for col in columns)) + "</tr>\n"
 
 
 @dataclasses.dataclass(slots=True, kw_only=True)
@@ -35,6 +43,12 @@ class Arguments(search.Arguments):
         ),
     )
     table_output: typing.Optional[TableOutput] = None
+    html: bool = dataclasses.field(
+        default=False,
+        metadata=argclasses.arg(
+            help="HTML table output - requires list of attributes",
+        ),
+    )
     sort: bool = dataclasses.field(
         default=False,
         metadata=argclasses.arg(
@@ -45,14 +59,16 @@ class Arguments(search.Arguments):
     def __post_init__(self) -> None:
         super(Arguments, self).__post_init__()  # super() not working here, unclear why.
 
-        # pick at most one in csv, (markdown) table
-        if [self.csv, self.table].count(True) > 1:
+        # pick at most one in csv, (markdown) table, html
+        if [self.csv, self.table, self.html].count(True) > 1:
             raise SystemExit("Can't use more than one table output type")
 
         if self.csv:
             self.table_output = TableOutput.CSV
         elif self.table:
             self.table_output = TableOutput.MARKDOWN
+        elif self.html:
+            self.table_output = TableOutput.HTML
 
         if self.sort and self.table_output is None:
             # default to markdown table
@@ -135,11 +151,20 @@ class _Context:
         if self.arguments.sort:
             line_iterator = sorted(line_iterator)
 
-        csv_out = csv.writer(stream, lineterminator="\n")
-        csv_out.writerow(self.arguments.columns)
+        if self.arguments.table_output in [TableOutput.CSV, TableOutput.MARKDOWN]:
+            csv_out = csv.writer(stream, lineterminator="\n")
+            csv_out.writerow(self.arguments.columns)
 
-        for line in line_iterator:
-            csv_out.writerow(line)
+            for line in line_iterator:
+                csv_out.writerow(line)
+        else:
+            assert self.arguments.table_output == TableOutput.HTML
+
+            stream.write("<table>\n")
+            stream.write(_html_escape_line(self.arguments.columns, cell="th"))
+            for line in line_iterator:
+                stream.write(_html_escape_line(line))
+            stream.write("</table>\n")
 
     def _ldif_or_json_output(self, search_iterator: typing.Iterable[Result], *, stream: typing.IO[str]) -> None:
         decoder = decode.Decoder(arguments=self.arguments)