diff --git a/modules/750g/pages.py b/modules/750g/pages.py
index 6f63b4b3e7a1bb77e8d4ddb2d812c7488d38e2b5..c81f0577d2d9025e707a070414a3d448082395ae 100644
--- a/modules/750g/pages.py
+++ b/modules/750g/pages.py
@@ -18,7 +18,7 @@
 # along with weboob. If not, see <http://www.gnu.org/licenses/>.
 
 
-from weboob.capabilities.recipe import Recipe
+from weboob.capabilities.recipe import Recipe, Comment
 from weboob.capabilities.base import NotAvailable, NotLoaded
 from weboob.tools.browser import BasePage
 
@@ -121,10 +121,13 @@ def get_recipe(self, id):
             comtxt = unicode(' '.join(divcom.text_content().strip().split()))
             if u'| Répondre' in comtxt:
                 comtxt = comtxt.strip('0123456789').replace(u' | Répondre', '')
-            comments.append(comtxt)
+                author = None
+                if 'par ' in comtxt:
+                    author = comtxt.split('par ')[-1].split('|')[0]
+                    comtxt = comtxt.replace('par %s' % author, '')
+            comments.append(Comment(text=comtxt, author=author))
             
         links_author = self.parser.select(self.document.getroot(), 'p.auteur a.couleur_membre')
-        print links_author[0].text.strip()
         if len(links_author) > 0:
             author = unicode(links_author[0].text.strip())
 
diff --git a/modules/cuisineaz/pages.py b/modules/cuisineaz/pages.py
index 38c37213270b09d618a209395288d82920dd9179..beda21d2f9581cb1e13a87752968b3d33fe5b97d 100644
--- a/modules/cuisineaz/pages.py
+++ b/modules/cuisineaz/pages.py
@@ -18,7 +18,7 @@
 # along with weboob. If not, see <http://www.gnu.org/licenses/>.
 
 
-from weboob.capabilities.recipe import Recipe
+from weboob.capabilities.recipe import Recipe, Comment
 from weboob.capabilities.base import NotAvailable, NotLoaded
 from weboob.tools.browser import BasePage
 
@@ -123,12 +123,9 @@ def get_recipe(self, id):
         for divcom in self.parser.select(self.document.getroot(), 'div.comment'):
             author = unicode(self.parser.select(
                 divcom, 'div.commentAuthor span', 1).text)
-            date = unicode(self.parser.select(
-                divcom, 'div.commentDate', 1).text)
             comtxt = unicode(self.parser.select(
                 divcom, 'p', 1).text_content().strip())
-            comments.append('author: %s, date: %s, text: %s' % (
-                author, date, comtxt))
+            comments.append(Comment(author=author, text=comtxt))
 
         spans_author = self.parser.select(self.document.getroot(), 'span.author')
         if len(spans_author) > 0:
diff --git a/modules/marmiton/pages.py b/modules/marmiton/pages.py
index 20812b23a1a9107079ff37a003dcbef6ae43af73..8ea6eaf3b59fdd2d0d359d090f3e5c15a620f83d 100644
--- a/modules/marmiton/pages.py
+++ b/modules/marmiton/pages.py
@@ -18,7 +18,7 @@
 # along with weboob. If not, see <http://www.gnu.org/licenses/>.
 
 
-from weboob.capabilities.recipe import Recipe
+from weboob.capabilities.recipe import Recipe, Comment
 from weboob.capabilities.base import NotAvailable, NotLoaded
 from weboob.tools.browser import BasePage
 
@@ -91,7 +91,7 @@ def get_recipe(self, id):
             note = self.parser.select(divcom, 'div.m_commentaire_note span', 1).text.strip()
             user = self.parser.select(divcom, 'div.m_commentaire_content span', 1).text.strip()
             content = self.parser.select(divcom, 'div.m_commentaire_content p', 1).text.strip()
-            comments.append(u'user: %s, note: %s, comment: %s' % (user, note, content))
+            comments.append(Comment(author=user, rate=note, text=content))
 
         recipe = Recipe(id, title)
         recipe.preparation_time = preparation_time
diff --git a/weboob/capabilities/recipe.py b/weboob/capabilities/recipe.py
index 88d612705ebab9c319d207f21daffa1d26b6cd02..261e06cd2b5bd4fab0027dfa0ce1606f12514198 100644
--- a/weboob/capabilities/recipe.py
+++ b/weboob/capabilities/recipe.py
@@ -26,6 +26,22 @@
 __all__ = ['Recipe', 'ICapRecipe']
 
 
+class Comment():
+    def __init__(self, author=None, rate=None, text=None):
+        self.author = author
+        self.rate = rate
+        self.text = text
+
+    def __str__(self):
+        result = u''
+        if self.author:
+            result += 'author: %s, ' % self.author
+        if self.rate:
+            result += 'note: %s, ' % self.rate
+        if self.text:
+            result += 'comment: %s' % self.text
+        return result
+
 class Recipe(CapBaseObject):
     """
     Recipe object.
@@ -117,17 +133,19 @@ def toKrecipesXml(self, author=None):
             ratings = ET.SubElement(recipe, 'krecipes-ratings')
             for c in self.comments:
                 rating = ET.SubElement(ratings, 'rating')
-                com = ET.SubElement(rating, 'comment')
-                com.text = c
+                if c.author:
+                    rater = ET.SubElement(rating, 'rater')
+                    rater.text = c.author
+                if c.text:
+                    com = ET.SubElement(rating, 'comment')
+                    com.text = c.text
                 crits = ET.SubElement(rating, 'criterion')
-                crit = ET.SubElement(crits, 'criteria')
-                critname = ET.SubElement(crit, 'name')
-                critname.text = ''
-                critstars = ET.SubElement(crit, 'stars')
-                critstars.text = ''
-
-                rater = ET.SubElement(rating, 'rater')
-                rater.text = ''
+                if c.rate:
+                    crit = ET.SubElement(crits, 'criteria')
+                    critname = ET.SubElement(crit, 'name')
+                    critname.text = 'Overall'
+                    critstars = ET.SubElement(crit, 'stars')
+                    critstars.text = c.rate.split('/')[0]
 
         return header + ET.tostring(doc, encoding='UTF-8', pretty_print=True).decode('utf-8')