diff --git a/graphify/rcfile.py b/graphify/rcfile.py
index 110c0eca1..d6a34091b 100644
--- a/graphify/rcfile.py
+++ b/graphify/rcfile.py
@@ -123,7 +123,10 @@ def load_graphifyrc(root: Path) -> dict:
 
     cfg: dict = {}
     languages: dict[str, str] = {}
-    content = rc_path.read_text(encoding="utf-8")
+    # utf-8-sig: Windows PowerShell 5.1 (`Set-Content -Encoding UTF8`) and older
+    # Notepad save UTF-8 with a BOM. Read as plain utf-8, the BOM became part of
+    # the first key, which then matched nothing and was ignored without a word.
+    content = rc_path.read_text(encoding="utf-8-sig")
     for line_num, raw in enumerate(content.splitlines(), 1):
         line = raw.strip()
         if not line or line.startswith("#"):
diff --git a/tests/test_language_overrides.py b/tests/test_language_overrides.py
index 9b38d398d..2bfbb5248 100644
--- a/tests/test_language_overrides.py
+++ b/tests/test_language_overrides.py
@@ -91,6 +91,19 @@ def test_language_line_maps_an_extension_to_a_canonical_suffix(tmp_path):
     assert load_graphifyrc(tmp_path) == {"languages": {".inc": ".php"}}
 
 
+def test_a_utf8_bom_does_not_swallow_the_first_key(tmp_path):
+    """A .graphifyrc saved with a UTF-8 BOM (Windows PowerShell 5.1, older
+    Notepad) must not glue the BOM onto its first key, which then matched
+    nothing and was dropped without a warning."""
+    (tmp_path / ".graphifyrc").write_bytes(
+        "\ufefflanguage.inc=php\nviz_node_limit=500\n".encode("utf-8")
+    )
+    assert load_graphifyrc(tmp_path) == {
+        "languages": {".inc": ".php"},
+        "viz_node_limit": 500,
+    }
+
+
 @pytest.mark.parametrize("key", ["language.inc", "language..inc", "language.INC", "language. inc "])
 def test_the_extension_key_is_normalised(tmp_path, key):
     (tmp_path / ".graphifyrc").write_text(f"{key}=php\n", encoding="utf-8")