diff --git a/blog/models.py b/blog/models.py
index bceb067d1..bc8844fac 100644
--- a/blog/models.py
+++ b/blog/models.py
@@ -1,3 +1,4 @@
+import xml.etree.ElementTree as etree
 from urllib.parse import urlparse
 
 from django.conf import settings
@@ -12,8 +13,11 @@
 from django.utils.translation import gettext_lazy as _
 from django_hosts.resolvers import get_host, reverse, reverse_host
 from docutils.core import publish_parts
-from markdown import markdown
+from docutils.nodes import document
+from docutils.writers.html4css1 import HTMLTranslator, Writer
+from markdown import Markdown
 from markdown.extensions.toc import TocExtension, slugify as _md_title_slugify
+from markdown.treeprocessors import Treeprocessor
 
 BLOG_DOCUTILS_SETTINGS = {
     "doctitle_xform": False,
@@ -38,6 +42,45 @@ def active(self):
         return self.filter(is_active=True)
 
 
+_IMG_LAZY_ATTRIBUTES = {"loading": "lazy"}
+
+
+class LazyImageHTMLTranslator(HTMLTranslator):
+    """Alter the img tags to include the lazy attribute."""
+
+    def __init__(
+        self, document: document, img_attributes: dict[str, str] | None = None
+    ) -> None:
+        super().__init__(document)
+        self._img_attributes = img_attributes or _IMG_LAZY_ATTRIBUTES
+
+    def emptytag(self, node, tagname, suffix="\n", **attributes):
+        """Construct and return an XML-compatible empty tag."""
+        if tagname == "img":
+            attributes.update(self._img_attributes)
+        return super().emptytag(node, tagname, suffix, **attributes)
+
+
+class LazyImageTreeprocessor(Treeprocessor):
+    """
+    `Treeprocessor`s are run on the `ElementTree` object before serialization.
+
+    This processor will add loading=lazy attribute on img tags
+
+    """
+
+    def __init__(
+        self, img_attributes: dict[str, str] | None = None, md: Markdown | None = None
+    ) -> None:
+        super().__init__(md)
+        self._img_attributes = img_attributes or _IMG_LAZY_ATTRIBUTES
+
+    def run(self, root: etree.Element) -> etree.Element | None:
+        """Alter img tags with the supplemental attributes."""
+        for img_elem in root.iter("img"):
+            img_elem.attrib.update(self._img_attributes)
+
+
 class ContentFormat(models.TextChoices):
     REST = "reST", "reStructuredText"
     HTML = "html", "Raw HTML"
@@ -51,20 +94,24 @@ def to_html(cls, fmt, source):
         if not fmt or fmt == cls.HTML:
             return source
         if fmt == cls.REST:
+            writer = Writer()
+            writer.translator_class = LazyImageHTMLTranslator
+
             return publish_parts(
                 source=source,
-                writer_name="html",
+                writer=writer,
                 settings_overrides=BLOG_DOCUTILS_SETTINGS,
             )["fragment"]
         if fmt == cls.MARKDOWN:
-            return markdown(
-                source,
+            md = Markdown(
                 output_format="html",
                 extensions=[
                     # baselevel matches `initial_header_level` from BLOG_DOCUTILS_SETTINGS
                     TocExtension(baselevel=3, slugify=_md_slugify),
                 ],
             )
+            md.treeprocessors.register(LazyImageTreeprocessor(), "lazyimage", 0.3)
+            return md.convert(source)
         raise ValueError(f"Unsupported format {fmt}")
 
     def img(self, url, alt_text):
diff --git a/blog/tests.py b/blog/tests.py
index c345fda56..dece5b19b 100644
--- a/blog/tests.py
+++ b/blog/tests.py
@@ -130,6 +130,33 @@ def test_header_base_level_markdown(self):
         )
         self.assertHTMLEqual(entry.body_html, '<h3 id="s-test">test</h3>')
 
+    def test_image_lazy_loader_markdown(self):
+        entry = Entry.objects.create(
+            pub_date=self.now,
+            slug="a",
+            body="# this is it\n![dilbert_alt](/m/blog/images/2025/10/Dilbert.gif)",
+            content_format=ContentFormat.MARKDOWN,
+        )
+        self.assertHTMLEqual(
+            entry.body_html,
+            '<h3 id="s-this-is-it">this is it</h3>'
+            '<p><img alt="dilbert_alt" loading="lazy"'
+            ' src="/m/blog/images/2025/10/Dilbert.gif"></p>',
+        )
+
+    def test_image_lazy_loader_reST(self):
+        entry = Entry.objects.create(
+            pub_date=self.now,
+            slug="a",
+            body=".. image:: /m/blog/images/2025/10/Dilbert.gif\n   :alt: dilbert_alt",
+            content_format=ContentFormat.REST,
+        )
+        self.assertHTMLEqual(
+            entry.body_html,
+            '<img alt="dilbert_alt" loading="lazy"'
+            ' src="/m/blog/images/2025/10/Dilbert.gif">',
+        )
+
     def test_pub_date_localized(self):
         entry = Entry(pub_date=date(2005, 7, 21))
         self.assertEqual(entry.pub_date_localized, "July 21, 2005")
@@ -581,33 +608,34 @@ def test_full_url(self):
         )
 
     def test_alt_text_html_escape(self):
+        common_img_tag = '<img src="." loading="lazy" '
         testdata = [
             (ContentFormat.HTML, 'te"st', '<img src="." alt="te&quot;st">'),
             (ContentFormat.HTML, "te<st>", '<img src="." alt="te&lt;st&gt;">'),
-            (ContentFormat.MARKDOWN, 'te"st', '<img src="." alt="te&quot;st">'),
-            (ContentFormat.MARKDOWN, "te[st]", '<img src="." alt="te[st]">'),
-            (ContentFormat.MARKDOWN, "te{st}", '<img src="." alt="te{st}">'),
-            (ContentFormat.MARKDOWN, "te<st>", '<img src="." alt="te&lt;st&gt;">'),
-            (ContentFormat.MARKDOWN, "test*", '<img src="." alt="test*">'),
-            (ContentFormat.MARKDOWN, "test_", '<img src="." alt="test_">'),
-            (ContentFormat.MARKDOWN, "test`", '<img src="." alt="test`">'),
-            (ContentFormat.MARKDOWN, "test+", '<img src="." alt="test+">'),
-            (ContentFormat.MARKDOWN, "test-", '<img src="." alt="test-">'),
-            (ContentFormat.MARKDOWN, "test.", '<img src="." alt="test.">'),
-            (ContentFormat.MARKDOWN, "test!", '<img src="." alt="test!">'),
-            (ContentFormat.MARKDOWN, "te\nst", '<img src="." alt="te\nst">'),
-            (ContentFormat.REST, 'te"st', '<img src="." alt="te&quot;st">'),
-            (ContentFormat.REST, "te[st]", '<img src="." alt="te[st]">'),
-            (ContentFormat.REST, "te{st}", '<img src="." alt="te{st}">'),
-            (ContentFormat.REST, "te<st>", '<img src="." alt="te&lt;st&gt;">'),
-            (ContentFormat.REST, "te:st", '<img src="." alt="te:st">'),
-            (ContentFormat.REST, "test*", '<img src="." alt="test*">'),
-            (ContentFormat.REST, "test_", '<img src="." alt="test_">'),
-            (ContentFormat.REST, "test`", '<img src="." alt="test`">'),
-            (ContentFormat.REST, "test+", '<img src="." alt="test+">'),
-            (ContentFormat.REST, "test-", '<img src="." alt="test-">'),
-            (ContentFormat.REST, "test.", '<img src="." alt="test.">'),
-            (ContentFormat.REST, "test!", '<img src="." alt="test!">'),
+            (ContentFormat.MARKDOWN, 'te"st', common_img_tag + 'alt="te&quot;st">'),
+            (ContentFormat.MARKDOWN, "te[st]", common_img_tag + 'alt="te[st]">'),
+            (ContentFormat.MARKDOWN, "te{st}", common_img_tag + 'alt="te{st}">'),
+            (ContentFormat.MARKDOWN, "te<st>", common_img_tag + 'alt="te&lt;st&gt;">'),
+            (ContentFormat.MARKDOWN, "test*", common_img_tag + 'alt="test*">'),
+            (ContentFormat.MARKDOWN, "test_", common_img_tag + 'alt="test_">'),
+            (ContentFormat.MARKDOWN, "test`", common_img_tag + 'alt="test`">'),
+            (ContentFormat.MARKDOWN, "test+", common_img_tag + 'alt="test+">'),
+            (ContentFormat.MARKDOWN, "test-", common_img_tag + 'alt="test-">'),
+            (ContentFormat.MARKDOWN, "test.", common_img_tag + 'alt="test.">'),
+            (ContentFormat.MARKDOWN, "test!", common_img_tag + 'alt="test!">'),
+            (ContentFormat.MARKDOWN, "te\nst", common_img_tag + 'alt="te\nst">'),
+            (ContentFormat.REST, 'te"st', common_img_tag + 'alt="te&quot;st">'),
+            (ContentFormat.REST, "te[st]", common_img_tag + 'alt="te[st]">'),
+            (ContentFormat.REST, "te{st}", common_img_tag + 'alt="te{st}">'),
+            (ContentFormat.REST, "te<st>", common_img_tag + 'alt="te&lt;st&gt;">'),
+            (ContentFormat.REST, "te:st", common_img_tag + 'alt="te:st">'),
+            (ContentFormat.REST, "test*", common_img_tag + 'alt="test*">'),
+            (ContentFormat.REST, "test_", common_img_tag + 'alt="test_">'),
+            (ContentFormat.REST, "test`", common_img_tag + 'alt="test`">'),
+            (ContentFormat.REST, "test+", common_img_tag + 'alt="test+">'),
+            (ContentFormat.REST, "test-", common_img_tag + 'alt="test-">'),
+            (ContentFormat.REST, "test.", common_img_tag + 'alt="test.">'),
+            (ContentFormat.REST, "test!", common_img_tag + 'alt="test!">'),
         ]
         for cf, alt_text, expected in testdata:
             # RST doesn't like an empty src, so we use . instead