diff --git a/.gitignore b/.gitignore
index 9263db3..0e9e459 100644
--- a/.gitignore
+++ b/.gitignore
@@ -9,3 +9,6 @@ __pycache__/
 
 # Persönliche Claude-Code-Einstellungen
 .claude/settings.local.json
+
+# Eigene Ergebnisse und Testläufe
+output/
diff --git a/aenderungen.py b/aenderungen.py
index 7c16a0f..3faf12f 100755
--- a/aenderungen.py
+++ b/aenderungen.py
@@ -11,9 +11,12 @@ import re
 
 from vtt2txt import vtt_to_lines
 
-REVIEW_MARK = re.compile(r"\[prüfen:[^\]]*\]")
-TOKEN = re.compile(r"\[prüfen:[^\]]*\]|\S+")
+# "[prüfen: …]" in German speeches, "[check: …]" in English ones
+REVIEW_MARK = re.compile(r"\[(?:prüfen|check):[^\]]*\]")
+TOKEN = re.compile(rf"{REVIEW_MARK.pattern}|\S+")
 NON_WORD = re.compile(r"[^\w]+")
+NON_WORD_KEEP_APOSTROPHE = re.compile(r"[^\w']+")
+APOSTROPHES = str.maketrans("’‘`´", "''''")
 ABBREVIATION = re.compile(r"\w\.")          # one part of "z. B.", "d. h."
 ABBREVIATION_END = re.compile(r"(?:^|\s)\w\.$")
 EMPTY = "∅"
@@ -32,8 +35,13 @@ def add_token(words, token):
 
 
 def normalize(word):
-    """Compare words without case and punctuation; those changes are expected."""
-    return NON_WORD.sub("", word.lower()) or word
+    """Compare words without case and punctuation; those changes are expected.
+
+    Apostrophes inside a word count ("Bahr’s" is not "Bahrs"), quotes at its
+    edges do not.
+    """
+    word = word.lower().translate(APOSTROPHES)
+    return NON_WORD_KEEP_APOSTROPHE.sub("", word).strip("'") or word
 
 
 @dataclass
diff --git a/tests/test_aenderungen.py b/tests/test_aenderungen.py
index ba01aaf..2dcd7d3 100644
--- a/tests/test_aenderungen.py
+++ b/tests/test_aenderungen.py
@@ -31,6 +31,26 @@ def test_expected_edits_are_no_changes(raw, corrected):
     assert changes(raw, corrected) == []
 
 
+@pytest.mark.parametrize("raw, corrected", [
+    ("gibt's", "gibt’s"),                                 # apostrophe variants
+    ("er sagte Frieden", "Er sagte ‚Frieden‘."),          # quotes at the edges
+])
+def test_apostrophes_and_quotes_unchanged(raw, corrected):
+    assert changes(raw, corrected) == []
+
+
+def test_apostrophe_inside_word_is_a_change():
+    assert changes("Egon Bahrs Politik", "Egon Bahr’s Politik") == [
+        ("replace", "Bahrs", "Bahr’s"),
+    ]
+
+
+def test_english_review_marks():
+    comparison = ae.compare("his name is Kocher", "His name is Kocher [check: Kocha].")
+    assert list(comparison.changes()) == []
+    assert sum(len(m) for m in comparison.marks.values()) == 1
+
+
 def test_replace_insert_delete():
     raw = "die gedeiliche Zusammenarbeit war das gut heute"
     corrected = "die gedeihliche Zusammenarbeit war doch gut"