diff --git a/src/codeweaver/engine/chunker/delimiter.py b/src/codeweaver/engine/chunker/delimiter.py
index 07b00994..0d8f4d90 100644
--- a/src/codeweaver/engine/chunker/delimiter.py
+++ b/src/codeweaver/engine/chunker/delimiter.py
@@ -451,6 +451,9 @@ def _match_keyword_delimiters(
         # Filter out delimiters with empty start strings - they match everywhere!
         keyword_delimiters = [d for d in keyword_delimiters if d.start]
 
+        if not keyword_delimiters:
+            return matches
+
         # Define structural delimiters that can complete keywords
         # Map opening structural chars to their closing counterparts
         structural_pairs = {
@@ -458,23 +461,33 @@ def _match_keyword_delimiters(
             ":": "\n",  # Python uses : followed by indented block (simplified to newline)
             "=>": "",  # Arrow functions often have expression bodies
         }
-
+        allowed_keys = frozenset(structural_pairs.keys())
+        # Optimization: Combine all keyword start strings into a single compiled regex pattern.
+        # This allows us to make a single pass over the content rather than iterating over
+        # `re.finditer` for each keyword delimiter individually, significantly reducing overhead.
+        start_strings = list(dict.fromkeys(d.start for d in keyword_delimiters))
+        combined_pattern = re.compile(rf"\b(?:{'|'.join(map(re.escape, start_strings))})\b")
+        
+        # Group delimiters by matched start string so duplicate keyword starts
+        # preserve the original "process each delimiter independently" behavior.
+        delimiter_map: dict[str, list[Delimiter]] = {}
         for delimiter in keyword_delimiters:
-            # Find all keyword occurrences using word boundary matching
-            pattern = rf"\b{re.escape(delimiter.start)}\b"
+            delimiter_map.setdefault(delimiter.start, []).append(delimiter)
 
-            for match in re.finditer(pattern, content):
-                keyword_pos = match.start()
+        for match in combined_pattern.finditer(content):
+            matched_text = match.group(0)
+            keyword_pos = match.start()
 
-                # Skip if keyword is inside a string or comment
-                if self._is_inside_string_or_comment(content, keyword_pos):
-                    continue
+            # Skip if keyword is inside a string or comment
+            if self._is_inside_string_or_comment(content, keyword_pos):
+                continue
 
+            for delimiter in delimiter_map[matched_text]:
                 # Find the next structural opening after the keyword
                 struct_start, struct_char = self._find_next_structural_with_char(
                     content,
                     start=keyword_pos + len(delimiter.start),
-                    allowed=set(structural_pairs.keys()),
+                    allowed=allowed_keys,
                 )
 
                 if struct_start is None:
@@ -1280,7 +1293,7 @@ def _load_custom_delimiters(
         patterns when merged with the full delimiter list.
 
         Args:
-            normalized_language: Snake-case normalised language identifier.
+            normalized_language: Snake-case normalized language identifier.
             language: Original language string (used for logging only).
 
         Returns:
diff --git a/tests/unit/engine/chunker/test_delimiter_edge_cases.py b/tests/unit/engine/chunker/test_delimiter_edge_cases.py
index ca800938..12b26439 100644
--- a/tests/unit/engine/chunker/test_delimiter_edge_cases.py
+++ b/tests/unit/engine/chunker/test_delimiter_edge_cases.py
@@ -16,7 +16,9 @@
 
 import pytest
 
+from codeweaver.core.types import DelimiterKind
 from codeweaver.engine import ChunkGovernor, DelimiterChunker
+from codeweaver.engine.chunker.delimiter_model import Delimiter
 
 
 pytestmark = [pytest.mark.unit]
@@ -208,6 +210,41 @@ def test_take_whole_lines_expansion(
 class TestUnusualPatterns:
     """Test delimiter chunker with unusual patterns."""
 
+    def test_duplicate_keyword_starts_preserve_all_delimiters(
+        self, delimiter_chunker: DelimiterChunker
+    ) -> None:
+        """Verify duplicate keyword starts keep all matching delimiters.
+
+        Keyword delimiters with the same start string used to be processed
+        independently. The optimized single-regex path must preserve that
+        behavior instead of collapsing duplicates down to the last delimiter.
+        """
+        content = """type Person struct {
+    name string
+}"""
+        keyword_delimiters = [
+            Delimiter(
+                start="type",
+                end="",
+                kind=DelimiterKind.STRUCT,
+                priority=DelimiterKind.STRUCT.default_priority,
+            ),
+            Delimiter(
+                start="type",
+                end="",
+                kind=DelimiterKind.TYPE_ALIAS,
+                priority=DelimiterKind.TYPE_ALIAS.default_priority,
+            ),
+        ]
+
+        matches = delimiter_chunker._match_keyword_delimiters(content, keyword_delimiters)
+
+        assert len(matches) == 2, "Duplicate keyword starts should preserve all delimiters"
+        assert {match.delimiter.kind for match in matches} == {
+            DelimiterKind.STRUCT,
+            DelimiterKind.TYPE_ALIAS,
+        }
+
     def test_nested_delimiter_structures(
         self, delimiter_chunker: DelimiterChunker, tmp_path: Path
     ) -> None: