diff --git a/README.md b/README.md
index e14dec69..ce55185f 100644
--- a/README.md
+++ b/README.md
@@ -216,6 +216,17 @@ How to change locale at runtime:
 '3 seconds ago'
 ```
 
+Number formatting also uses the locale's configured numeric separators:
+
+```pycon
+>>> _t = humanize.i18n.activate("tr_TR")
+>>> humanize.intcomma(1_234_567.89)
+'1.234.567,89'
+>>> humanize.intword(1_200_000)
+'1,2 milyon'
+>>> humanize.i18n.deactivate()
+```
+
 You can pass additional parameter `path` to `activate` to specify a path to search
 locales in.
 
diff --git a/src/humanize/i18n.py b/src/humanize/i18n.py
index 0bf73de5..672dddc5 100644
--- a/src/humanize/i18n.py
+++ b/src/humanize/i18n.py
@@ -26,6 +26,7 @@
     "pt_BR": ".",
     "hu_HU": " ",
     "lv": " ",
+    "tr_TR": ".",
 }
 
 # Mapping of locale to decimal separator
@@ -36,6 +37,7 @@
     "pt_BR": ",",
     "hu_HU": ",",
     "lv": ",",
+    "tr_TR": ",",
 }
 
 
diff --git a/tests/test_i18n.py b/tests/test_i18n.py
index 683fe3ce..d75ed1ef 100644
--- a/tests/test_i18n.py
+++ b/tests/test_i18n.py
@@ -124,6 +124,12 @@ def test_intcomma() -> None:
         humanize.i18n.activate("pt_BR")
         assert humanize.intcomma(number) == "10.000.000"
 
+        humanize.i18n.activate("tr_TR")
+        assert humanize.intcomma(number) == "10.000.000"
+        assert humanize.intcomma(1_234_567.89) == "1.234.567,89"
+        assert humanize.intcomma("1234567,89") == "1.234.567,89"
+        assert humanize.intcomma("1.234.567,89") == "1.234.567,89"
+
     except FileNotFoundError:
         pytest.skip("Generate .mo with scripts/generate-translation-binaries.sh")
 
@@ -159,6 +165,9 @@ def test_naturaldelta() -> None:
         ("it_IT", 1_200_000, "1,2 milioni"),
         ("it_IT", 1_000_000_000, "1,0 miliardo"),
         ("it_IT", 3_500_000_000, "3,5 miliardi"),
+        # Turkish uses comma as decimal separator
+        ("tr_TR", 1_000_000, "1,0 milyon"),
+        ("tr_TR", 1_200_000, "1,2 milyon"),
         # Sinhala uses borrowed scale words with a dot decimal separator
         ("si_LK", 1_000_000, "1.0 මිලියන"),
         ("si_LK", 1_200_000, "1.2 මිලියන"),