diff --git a/shelfmark/release_sources/audiobookbay/scraper.py b/shelfmark/release_sources/audiobookbay/scraper.py
index dc6c7597..ff35643b 100644
--- a/shelfmark/release_sources/audiobookbay/scraper.py
+++ b/shelfmark/release_sources/audiobookbay/scraper.py
@@ -1,8 +1,10 @@
 """Web scraping functions for AudiobookBay."""
 
+import base64
 import re
 import time
 from threading import Lock
+from typing import TYPE_CHECKING
 from urllib.parse import quote, quote_plus
 
 import requests
@@ -14,6 +16,9 @@ from shelfmark.download import http as downloader
 from shelfmark.download.postprocess.packs import PackFile
 from shelfmark.release_sources.audiobookbay.utils import normalize_search_punctuation
 
+if TYPE_CHECKING:
+    from bs4.element import Tag
+
 logger = setup_logger(__name__)
 
 # Default trackers if none found on page
@@ -119,6 +124,26 @@ def _encode_search_query(query: str, *, exact_phrase: bool) -> str:
     return quote_plus(search_query)
 
 
+def _decode_post(post: Tag) -> Tag | BeautifulSoup:
+    """Decode ABB's base64-encoded post markup when present.
+
+    ABB serves post content base64-encoded inside a hidden ``.post`` div and
+    decodes it client-side with JavaScript; the decoded payload is the classic
+    ``.postTitle``/``.postInfo``/``.postContent`` markup. Plain-markup posts
+    (older mirrors) are returned unchanged, so never assume base64.
+    """
+    raw = "".join(post.get_text(strip=True).split())
+    if not raw or len(raw) < 32:
+        return post
+    try:
+        html = base64.b64decode(raw, validate=True).decode("utf-8", "replace")
+    except ValueError:
+        return post
+    if "postTitle" not in html:
+        return post
+    return BeautifulSoup(html, "html.parser")
+
+
 def _normalize_result_url(url: str, hostname: str) -> str:
     """Normalize ABB result URLs to absolute HTTPS URLs."""
     normalized_url = (url or "").strip()
@@ -249,12 +274,13 @@ def search_audiobookbay(
             # Parse HTML
             soup = BeautifulSoup(page_html, "html.parser")
 
-            # Extract book entries
-            posts = soup.select(".post")
+            # Extract book entries (decoding ABB's base64-wrapped markup when present)
+            posts = [_decode_post(p) for p in soup.select(".post")]
             if not posts:
                 # No more results
                 break
 
+            parsed_before = len(results)
             for post in posts:
                 try:
                     # Extract title
@@ -351,6 +377,16 @@ def search_audiobookbay(
                     logger.debug("Skipping post due to error: %s", e)
                     continue
 
+            # A page with .post elements that all fail to parse usually means ABB
+            # changed its markup; without this the source silently returns [].
+            if len(results) == parsed_before:
+                logger.warning(
+                    "ABB page %s: found %d .post elements but parsed 0 results, "
+                    "site markup may have changed",
+                    page,
+                    len(posts),
+                )
+
             # Rate limiting delay between pages
             if page < max_pages and rate_limit_delay > 0:
                 time.sleep(rate_limit_delay)
diff --git a/tests/audiobookbay/test_scraper.py b/tests/audiobookbay/test_scraper.py
index e9191b58..aa851bc7 100644
--- a/tests/audiobookbay/test_scraper.py
+++ b/tests/audiobookbay/test_scraper.py
@@ -2,6 +2,7 @@
 Tests for AudiobookBay scraper functions.
 """
 
+import base64
 from unittest.mock import patch
 
 from shelfmark.release_sources.audiobookbay import scraper
@@ -51,6 +52,41 @@ EMPTY_SEARCH_HTML = """
 </html>
 """
 
+# ABB serves post markup base64-encoded inside hidden .post divs, decoded
+# client-side by JavaScript. The decoded payload is the classic markup.
+_POST_INNER_HTML = (
+    '<div class="postTitle"><h2><a href="/abss/encoded-book-title/" rel="bookmark">'
+    "Encoded Book Title - Encoded Author</a></h2></div>"
+    '<div class="postInfo">Category: Crime&nbsp; <br>Language: English'
+    '<span style="margin-left:100px;">Keywords: Test&nbsp;</span><br></div>'
+    '<div class="postContent">'
+    '<div class="center">'
+    '<p class="center"><a href="https://audiobookbay.lu/abss/encoded-book-title/">'
+    '<img src="https://example.com/encoded-cover.jpg" alt="Cover" width="250"></a></p>'
+    "</div>"
+    '<p style="text-align:center;">Posted: 12 Jul 2026<br>'
+    'Format: <span style="color:#a00;">M4B</span> / '
+    'Bitrate: <span style="color:#a00;">64 Kbps</span><br>'
+    'File Size: <span style="color:#00f;">1.55</span> GBs</p>'
+    "</div>"
+)
+
+_BASE64_PAYLOAD = base64.b64encode(_POST_INNER_HTML.encode("utf-8")).decode("ascii")
+
+PLAIN_POST_SEARCH_HTML = f'<html><body><div class="post">{_POST_INNER_HTML}</div></body></html>'
+
+BASE64_SEARCH_HTML = f'<html><body><div class="post re-ab" style="display:none;">{_BASE64_PAYLOAD}</div></body></html>'
+
+# .post elements whose content is neither valid base64 nor classic markup.
+UNPARSEABLE_SEARCH_HTML = """
+<html>
+<body>
+<div class="post re-ab" style="display:none;">this is not base64 and has no postTitle markup!</div>
+<div class="post re-ab" style="display:none;">another opaque blob without any known structure</div>
+</body>
+</html>
+"""
+
 # Mock HTML for detail page with info hash and trackers
 SAMPLE_DETAIL_HTML = """
 <html>
@@ -381,6 +417,56 @@ class TestSearchAudiobookbay:
         )
 
 
+class TestSearchAudiobookbayBase64Posts:
+    """Tests for ABB's base64-encoded post markup."""
+
+    @staticmethod
+    def _search(html):
+        with (
+            patch("shelfmark.release_sources.audiobookbay.scraper.downloader.html_get_page") as get,
+            patch("shelfmark.release_sources.audiobookbay.scraper.config.get") as cfg,
+        ):
+            cfg.return_value = 0.0
+            get.return_value = (html, "https://audiobookbay.lu/?s=test&cat=undefined%2Cundefined")
+            return scraper.search_audiobookbay("test", max_pages=1, hostname="audiobookbay.lu")
+
+    def test_base64_page_parses_same_posts_as_plain_page(self):
+        plain = self._search(PLAIN_POST_SEARCH_HTML)
+        encoded = self._search(BASE64_SEARCH_HTML)
+
+        assert len(encoded) == 1
+        assert encoded == plain
+        assert encoded[0]["title"] == "Encoded Book Title - Encoded Author"
+        assert encoded[0]["link"] == "https://audiobookbay.lu/abss/encoded-book-title/"
+        assert encoded[0]["format"] == "M4B"
+        assert encoded[0]["bitrate"] == "64 Kbps"
+        assert encoded[0]["size"] == "1.55 GB"
+        assert encoded[0]["posted_date"] == "12 Jul 2026"
+        assert encoded[0]["cover"] == "https://example.com/encoded-cover.jpg"
+
+    def test_warns_when_posts_found_but_none_parsed(self):
+        # setup_logger builds parentless loggers (no propagation), so caplog
+        # can't see them; spy on the module logger instead.
+        with patch.object(scraper.logger, "warning") as mock_warning:
+            results = self._search(UNPARSEABLE_SEARCH_HTML)
+
+        assert results == []
+        warning_texts = [
+            (call.args[0] % tuple(call.args[1:])) if len(call.args) > 1 else call.args[0]
+            for call in mock_warning.call_args_list
+        ]
+        assert any("2 .post elements" in t and "parsed 0" in t for t in warning_texts), (
+            warning_texts
+        )
+
+    def test_no_warning_when_posts_parse(self):
+        with patch.object(scraper.logger, "warning") as mock_warning:
+            results = self._search(SAMPLE_SEARCH_HTML)
+
+        assert results
+        mock_warning.assert_not_called()
+
+
 class TestExtractMagnetLink:
     """Tests for the extract_magnet_link function."""