diff --git a/apkg_builder.py b/apkg_builder.py
index 6599026..af00216 100644
--- a/apkg_builder.py
+++ b/apkg_builder.py
@@ -190,7 +190,8 @@ CARD_CSS = """
   direction: rtl;
   text-align: right;
   font-style: italic;
-  margin-top: 10px;
+  margin: 10px auto 0;
+  max-width: 90%;
   border-right: 3px solid #aaa;
   padding-right: 8px;
 }
@@ -211,12 +212,16 @@ CARD_CSS = """
   color: #555;
 }
 .sec-label {
-  font-size: 20px;
+  font-size: 32px;
   color: #555;
   direction: rtl;
   text-align: center;
   margin-top: 6px;
 }
+.sec-key {
+  font-size: 24px;
+  color: #888;
+}
 .related-group {
   direction: rtl;
   text-align: right;
@@ -235,6 +240,7 @@ CARD_CSS = """
   .meaning     { color: #82b0ff; }
   .root-info   { color: #aaa; }
   .sec-label   { color: #aaa; }
+  .sec-key     { color: #666; }
   .hint        { color: #777; }
   .voice-label { color: #888; }
   .example     { color: #bbb; border-right-color: #555; }
@@ -257,14 +263,14 @@ VOCAB_BACK_HEB = """
 <div class="divider"></div>
 <div class="meaning">{{Meaning}}</div>
 {{#Emoji}}<div class="emoji-img">{{Emoji}}</div>{{/Emoji}}
-{{^Emoji}}{{#Image}}<div><img src="{{Image}}" style="max-height:150px;margin-top:8px;"></div>{{/Image}}{{/Emoji}}
-{{#Root}}<div class="sec-label">שורש: {{Root}}</div>{{/Root}}
-{{#PoS}}<div class="sec-label">חלק דיבור: {{PoS}}</div>{{/PoS}}
+{{^Emoji}}{{#Image}}<div><img src="{{Image}}" style="max-height:150px;margin-top:8px;" onerror="this.parentElement.style.display='none'"></div>{{/Image}}{{/Emoji}}
+{{#Root}}<div class="sec-label"><span class="sec-key">שֹׁרֶשׁ:</span> {{Root}}</div>{{/Root}}
+{{#PoS}}<div class="sec-label"><span class="sec-key">חֵלֶק דִּיבּוּר:</span> {{PoS}}</div>{{/PoS}}
 {{#SharedRoots}}
-<div class="sec-label">מילים קשורות:</div>
+<div class="sec-label"><span class="sec-key">מִילִים קְשׁוּרוֹת:</span></div>
 <div class="root-info">{{SharedRoots}}</div>
 {{/SharedRoots}}
-{{#Plural}}<div class="sec-label">רבים: <span class="hebrew-sm">{{Plural}}</span></div>{{/Plural}}
+{{#Plural}}<div class="sec-label"><span class="sec-key">רַבִּים:</span> {{Plural}}</div>{{/Plural}}
 {{#Example}}
 <div class="example">{{Example}}</div>
 {{/Example}}
@@ -275,7 +281,7 @@ VOCAB_FRONT_ENG = """
 <div class="meaning">{{Meaning}}</div>
 {{#Hint}}<div class="hint">{{Hint}}</div>{{/Hint}}
 {{#Emoji}}<div class="emoji-img">{{Emoji}}</div>{{/Emoji}}
-{{^Emoji}}{{#Image}}<div><img src="{{Image}}" style="max-height:150px;margin-top:8px;"></div>{{/Image}}{{/Emoji}}
+{{^Emoji}}{{#Image}}<div><img src="{{Image}}" style="max-height:150px;margin-top:8px;" onerror="this.parentElement.style.display='none'"></div>{{/Image}}{{/Emoji}}
 """
 
 VOCAB_BACK_ENG = """
@@ -283,12 +289,10 @@ VOCAB_BACK_ENG = """
 <div class="divider"></div>
 <div class="hebrew">{{Word}}{{#Prep}} <span class="hebrew-sm">{{Prep}}</span>{{/Prep}}</div>
 {{#Audio}}<div>{{Audio}}</div>{{/Audio}}
-{{#WordNoNikkud}}<div class="sec-label">ללא ניקוד: {{WordNoNikkud}}</div>{{/WordNoNikkud}}
-{{#Root}}<div class="sec-label">שורש: {{Root}}</div>{{/Root}}
-{{#PoS}}<div class="sec-label">חלק דיבור: {{PoS}}</div>{{/PoS}}
-{{#Emoji}}<div class="emoji-img">{{Emoji}}</div>{{/Emoji}}
-{{^Emoji}}{{#Image}}<div><img src="{{Image}}" style="max-height:150px;margin-top:8px;"></div>{{/Image}}{{/Emoji}}
-{{#Plural}}<div class="sec-label">רבים: <span class="hebrew-sm">{{Plural}}</span></div>{{/Plural}}
+{{#WordNoNikkud}}<div class="sec-label"><span class="sec-key">לְלֹא נִיקּוּד:</span> {{WordNoNikkud}}</div>{{/WordNoNikkud}}
+{{#Root}}<div class="sec-label"><span class="sec-key">שֹׁרֶשׁ:</span> {{Root}}</div>{{/Root}}
+{{#PoS}}<div class="sec-label"><span class="sec-key">חֵלֶק דִּיבּוּר:</span> {{PoS}}</div>{{/PoS}}
+{{#Plural}}<div class="sec-label"><span class="sec-key">רַבִּים:</span> {{Plural}}</div>{{/Plural}}
 {{#Example}}
 <div class="example">{{Example}}</div>
 {{/Example}}
@@ -449,7 +453,6 @@ VOICE_MAP = {
 # ──────────────────────────────────────────────────────────────────────────────
 
 
-
 def _audio_tag(word_no_nikkud: str, audio_dir: Path = AUDIO_DIR) -> str:
     """Return [sound:xxx.mp3] if audio file exists, else empty string."""
     safe = re.sub(r"[^\u05d0-\u05ea]", "", word_no_nikkud)
@@ -738,6 +741,7 @@ def build_vocab_deck(
 
     # Load noun plural forms for vocab card back display
     noun_plural_lookup: dict[str, str] = {}  # word (nikkud) → plural (nikkud)
+    _noun_plural_stripped: dict[str, str] = {}  # word (stripped) → plural (nikkud), fallback
     noun_plural_path = DATA_DIR / "noun_plurals.json"
     if noun_plural_path.exists():
         try:
@@ -748,6 +752,9 @@ def build_vocab_deck(
                 pl = _entry.get("plural", "")
                 if sg and pl:
                     noun_plural_lookup[sg] = pl
+                    s = _strip_nikkud(sg)
+                    if s not in _noun_plural_stripped:
+                        _noun_plural_stripped[s] = pl
             logger.info(f"  Noun plurals loaded: {len(noun_plural_lookup)} entries")
         except (json.JSONDecodeError, OSError):
             pass
@@ -933,21 +940,25 @@ def build_vocab_deck(
             if mp3_path not in media_files:
                 media_files.append(mp3_path)
 
+        # Consonant-only form for confusable detection and cloze matching
+        word_consonants = _strip_nikkud(word)
+        is_confusable = word_consonants in _confusable_words
+
         # Example sentences — priority: EPUB (nikkud'd) > Ben Yehuda > none
+        # For confusable words (same consonants, different nikkud), only match by
+        # exact nikkud form to avoid showing wrong-word sentences.
         example_html = ""
         # 1. EPUB/PDF sentences (full nikkud)
-        epub_sents = (
-            epub_examples.get(word) or epub_examples.get(word_no_nik) or epub_examples.get(_strip_nikkud(word_no_nik))
-        )
+        epub_sents = epub_examples.get(word)
+        if not epub_sents and not is_confusable:
+            epub_sents = epub_examples.get(word_no_nik) or epub_examples.get(_strip_nikkud(word_no_nik))
         if epub_sents:
             example_html = epub_sents[0]
         else:
             # 2. Ben Yehuda examples (some have nikkud from nikkud corpus)
-            by_sents = (
-                examples_cache.get(word)
-                or examples_cache.get(word_no_nik)
-                or examples_cache.get(_strip_nikkud(word_no_nik))
-            )
+            by_sents = examples_cache.get(word)
+            if not by_sents and not is_confusable:
+                by_sents = examples_cache.get(word_no_nik) or examples_cache.get(_strip_nikkud(word_no_nik))
             if by_sents:
                 # Prefer nikkud'd Ben Yehuda sentences (contain combining marks)
                 nikkud_sents = [s for s in by_sents if any("\u0591" <= c <= "\u05c7" for c in s)]
@@ -958,13 +969,12 @@ def build_vocab_deck(
         # Uses stripped (no-nikkud) matching. Skips homographs (confusable words).
         cloze_example = ""
         cloze_hint = ""
-        word_consonants = _strip_nikkud(word)
-        if word_consonants and word_consonants not in _confusable_words:
+        if word_consonants and not is_confusable:
             # Pick best sentence for cloze: vetted first, then example_html
             cloze_source = None
-            vetted = (
-                vetted_cloze.get(word) or vetted_cloze.get(word_no_nik) or vetted_cloze.get(_strip_nikkud(word_no_nik))
-            )
+            vetted = vetted_cloze.get(word)
+            if not vetted and not is_confusable:
+                vetted = vetted_cloze.get(word_no_nik) or vetted_cloze.get(_strip_nikkud(word_no_nik))
             if vetted:
                 cloze_source = vetted[0]
             elif example_html:
@@ -1041,7 +1051,7 @@ def build_vocab_deck(
                 emoji_str,
                 prep_str,
                 hint_str,
-                noun_plural_lookup.get(word, ""),
+                noun_plural_lookup.get(word, "") or _noun_plural_stripped.get(word_consonants, ""),
                 cloze_example,
                 cloze_hint,
             ],