diff --git a/klausur-service/backend/cv_vocab_pipeline.py b/klausur-service/backend/cv_vocab_pipeline.py
index 77ebf50..88075bb 100644
--- a/klausur-service/backend/cv_vocab_pipeline.py
+++ b/klausur-service/backend/cv_vocab_pipeline.py
@@ -4273,14 +4273,22 @@ def _fix_phonetic_brackets(
     if not IPA_AVAILABLE:
         return entries
 
+    replaced_count = 0
     for entry in entries:
         for field in ('english', 'german', 'example'):
             text = entry.get(field, '') or ''
             # Check for any bracket type — Tesseract garbles [ into { or (
             if not any(ch in text for ch in '[{('):
                 continue
-            entry[field] = _replace_phonetics_in_text(text, pronunciation)
+            new_text = _replace_phonetics_in_text(text, pronunciation)
+            if new_text != text:
+                logger.info(f"_fix_phonetic_brackets: {field} '{text}' → '{new_text}'")
+                replaced_count += 1
+            else:
+                logger.info(f"_fix_phonetic_brackets: no change for {field} '{text}'")
+            entry[field] = new_text
 
+    logger.info(f"_fix_phonetic_brackets: {replaced_count} replacements in {len(entries)} entries")
     return entries
 
 
@@ -4325,21 +4333,26 @@ def _replace_phonetics_in_text(text: str, pronunciation: str = 'british') -> str
     def replacer(match):
         word = match.group(1)
         bracket_content = match.group(2).strip()
+        full_match = match.group(0)
 
         # Skip if bracket content looks like regular text (multiple words)
         if len(bracket_content.split()) > 3:
-            return match.group(0)
+            logger.info(f"  phonetic replacer: SKIP (too many words) '{full_match}'")
+            return full_match
 
         # Skip if bracket content is a meaningful word/prefix — e.g. (zer)brechen,
         # Tanz(veranstaltung). These are real German morphemes, not garbled IPA.
         if _is_meaningful_bracket_content(bracket_content):
-            return match.group(0)
+            logger.info(f"  phonetic replacer: SKIP (meaningful content) '{full_match}'")
+            return full_match
 
         # Look up in IPA dictionary
         ipa = _lookup_ipa(word, pronunciation)
         if not ipa:
-            return match.group(0)  # Keep original
+            logger.info(f"  phonetic replacer: SKIP (no IPA for '{word}') '{full_match}'")
+            return full_match  # Keep original
 
+        logger.info(f"  phonetic replacer: REPLACE '{full_match}' → '{word} [{ipa}]'")
         return f"{word} [{ipa}]"
 
     return _PHONETIC_BRACKET_RE.sub(replacer, text)