diff --git a/extruct/w3cmicrodata.py b/extruct/w3cmicrodata.py
index a254a5c6..a9e1c507 100644
--- a/extruct/w3cmicrodata.py
+++ b/extruct/w3cmicrodata.py
@@ -11,6 +11,7 @@
 
 import collections
 from functools import partial
+from copy import deepcopy
 
 try:
     from urlparse import urljoin
@@ -18,11 +19,32 @@
     from urllib.parse import urljoin
 
 import lxml.etree
+from lxml.html.clean import Cleaner
 from w3lib.html import strip_html5_whitespace
+import html_text
 
 from extruct.utils import parse_html
 
 
+# Cleaner which is similar to html_text cleaner, but is less aggressive
+cleaner = Cleaner(
+    scripts=True,
+    javascript=False,  # onclick attributes are fine
+    comments=True,
+    style=True,
+    links=True,
+    meta=True,
+    page_structure=False,  # <title> may be nice to have
+    processing_instructions=True,
+    embedded=False,  # keep embedded content
+    frames=False,  # keed frames
+    forms=False,  # keep forms
+    annoying_tags=False,
+    remove_unknown_tags=False,
+    safe_attrs_only=False,
+)
+
+
 class LxmlMicrodataExtractor(object):
     _xp_item = lxml.etree.XPath('descendant-or-self::*[@itemscope]')
     _xp_prop = lxml.etree.XPath("""set:difference(.//*[@itemprop],
@@ -49,11 +71,12 @@ def extract(self, htmlstring, base_url=None, encoding="UTF-8"):
         return self.extract_items(tree, base_url)
 
     def extract_items(self, document, base_url):
+        cleaned_document = cleaner.clean_html(document)
         items_seen = set()
         return [
             item for item in (
                 self._extract_item(it, items_seen=items_seen, base_url=base_url)
-                for it in self._xp_item(document))
+                for it in self._xp_item(cleaned_document))
             if item]
 
     def _extract_item(self, node, items_seen, base_url):
@@ -182,7 +205,7 @@ def _extract_property_value(self, node, items_seen, base_url, force=False):
             return self._extract_textContent(node)
 
     def _extract_textContent(self, node):
-        return u"".join(self._xp_clean_text(node)).strip()
+        return html_text.etree_to_text(node)
 
 
 MicrodataExtractor = LxmlMicrodataExtractor
diff --git a/requirements.txt b/requirements.txt
index 87a27224..04bc5c16 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -9,3 +9,4 @@ rdflib-jsonld
 mf2py>=1.1.0
 six
 w3lib
+html-text
diff --git a/setup.py b/setup.py
index 5bbc7553..f7e60387 100644
--- a/setup.py
+++ b/setup.py
@@ -33,6 +33,7 @@ def get_version():
                       'rdflib-jsonld', 
                       'mf2py', 
                       'w3lib',
+                      'html-text>=0.5.1',
                       'six'],
     extras_require={
         'service': [