diff --git a/docs/formatting.md b/docs/formatting.md
index 8387a731..1368356b 100644
--- a/docs/formatting.md
+++ b/docs/formatting.md
@@ -201,6 +201,12 @@ Conversion specifiers allow to *convert* the value to a different form or type.
     <td><code>{html!H}</code></td>
     <td><code>foo &amp; bar</code></td>
 </tr>
+<tr>
+    <td align="center"><code>R</code></td>
+    <td>Extract URLs</td>
+    <td><code>{lorem!R}</code></td>
+    <td><code>["https://example.org/"]</code></td>
+</tr>
 <tr>
     <td align="center"><code>s</code></td>
     <td>Convert value to <a href="https://docs.python.org/3/library/stdtypes.html#text-sequence-type-str" rel="nofollow"><code>str</code></a></td>
diff --git a/gallery_dl/formatter.py b/gallery_dl/formatter.py
index 7a49049b..b09203fb 100644
--- a/gallery_dl/formatter.py
+++ b/gallery_dl/formatter.py
@@ -565,6 +565,7 @@ _CONVERSIONS = {
     "U": text.unescape,
     "H": lambda s: text.unescape(text.remove_html(s)),
     "g": text.slugify,
+    "R": text.re(r"https?://[^\s\"']+").findall,
     "W": text.sanitize_whitespace,
     "S": util.to_string,
     "s": str,
diff --git a/test/test_formatter.py b/test/test_formatter.py
index f3ed9dd3..8b35a2bd 100644
--- a/test/test_formatter.py
+++ b/test/test_formatter.py
@@ -44,6 +44,10 @@ class TestFormatter(unittest.TestCase):
         "s": " \n\r\tSPACE    ",
         "S": " \n\r\tS  P         A\tC\nE    ",
         "h": "<p>foo </p> &amp; bar <p> </p>",
+        "H": """<p>
+  <a href="http://www.example.com">Lorem ipsum dolor sit amet</a>.
+  Duis aute irure <a href="http://blog.example.org">dolor</a>.
+</p>""",
         "u": "&#x27;&lt; / &gt;&#x27;",
         "t": 1262304000,
         "ds": "2010-01-01T01:00:00+01:00",
@@ -72,6 +76,9 @@ class TestFormatter(unittest.TestCase):
         self._run_test("{h!H}", "foo & bar")
         self._run_test("{u!H}", "'< / >'")
         self._run_test("{n!H}", "")
+        self._run_test("{h!R}", [])
+        self._run_test("{H!R}", ["http://www.example.com",
+                                 "http://blog.example.org"])
         self._run_test("{a!s}", self.kwdict["a"])
         self._run_test("{a!r}", f"'{self.kwdict['a']}'")
         self._run_test("{a!a}", f"'{self.kwdict['a']}'")
@@ -590,10 +597,11 @@ def gentext(kwdict):
 def lengths(kwdict):
     a = 0
     for k, v in kwdict.items():
-        try:
-            a += len(v)
-        except TypeError:
-            pass
+        if k == k.lower():
+            try:
+                a += len(v)
+            except TypeError:
+                pass
     return format(a)
 
 def noarg():
@@ -616,10 +624,10 @@ def noarg():
             fmt4 = formatter.parse(f"\fM {path}:lengths")
 
         self.assertEqual(fmt1.format_map(self.kwdict), "'Title' by Name")
-        self.assertEqual(fmt2.format_map(self.kwdict), "168")
+        self.assertEqual(fmt2.format_map(self.kwdict), "139")
 
         self.assertEqual(fmt3.format_map(self.kwdict), "'Title' by Name")
-        self.assertEqual(fmt4.format_map(self.kwdict), "168")
+        self.assertEqual(fmt4.format_map(self.kwdict), "139")
 
         with self.assertRaises(TypeError):
             self.assertEqual(fmt0.format_map(self.kwdict), "")