diff --git a/.jules/bolt.md b/.jules/bolt.md
index 341c7c91..2da6e608 100644
--- a/.jules/bolt.md
+++ b/.jules/bolt.md
@@ -67,3 +67,6 @@
 ## 2025-02-12 - [Fast Path Execution in Directory Traversal and Log Parsing]
 **Learning:** Checking for string existence (`if "silence_" not in stderr`) before invoking regex matchers provides significant speed improvements when parsing large blocks of text. Similarly, moving expensive I/O operations like `os.path.realpath` inside conditional blocks prevents redundant disk access when configuration (like path exclusions) isn't utilized.
 **Action:** When working on large text processing or disk operations, verify if early exit conditions or conditional execution can bypass the expensive system or library calls.
+## 2026-06-25 - [Fast Path for Alphanumeric Regex Replacements]
+**Learning:** Using regex substitution (e.g. `re.sub(r"^\W+|\W+$", "")`) in tight tokenization loops is extremely CPU-bound and slow. However, the majority of tokens in standard text are purely alphanumeric without any punctuation.
+**Action:** When tokenizing or processing text using regular expressions to strip non-word characters, prepend a fast-path check using `str.isalnum()` (or `str.isalpha()`) to bypass the regex overhead for purely alphanumeric words. This safely yields significant performance gains for most valid text inputs.
diff --git a/summarize.py b/summarize.py
index a5b56fe0..81e6b4e9 100644
--- a/summarize.py
+++ b/summarize.py
@@ -94,6 +94,14 @@ def _content_words(sentence):
     """
     words = []
     for raw in sentence.split():
+        # Fast path check: if raw is alphanumeric, skip regex overhead.
+        # This significantly improves performance because most tokens
+        # do not contain punctuation and safely bypass the regex substitution.
+        if raw.isalnum():
+            token = raw.lower()
+            if token not in _STOPWORDS:
+                words.append(token)
+            continue
         token = _TOKEN_STRIP_RE.sub("", raw).lower()
         if token and token not in _STOPWORDS:
             words.append(token)