diff --git a/pagure/lib/encoding_utils.py b/pagure/lib/encoding_utils.py
index 00c47bb..60c1493 100644
--- a/pagure/lib/encoding_utils.py
+++ b/pagure/lib/encoding_utils.py
@@ -15,7 +15,7 @@ from __future__ import unicode_literals, division, absolute_import
 from collections import namedtuple
 import logging
 
-from chardet import universaldetector
+from chardet import universaldetector, __version__ as ch_version
 
 from pagure.exceptions import PagureEncodingException
 
@@ -51,9 +51,17 @@ def detect_encodings(data):
     if not result:
         return {'utf-8': 1.0}
     encodings = {result['encoding']: result['confidence']}
-    for prober in detector._mCharSetProbers:
-        if prober:
-            encodings[prober.get_charset_name()] = prober.get_confidence()
+    if ch_version[0] == '3':
+        for prober in detector._charset_probers:
+            if hasattr(prober, 'probers'):
+                for prober in prober.probers:
+                    encodings[prober.charset_name] = prober.get_confidence()
+            else:
+                encodings[prober.charset_name] = prober.get_confidence()
+    else:
+        for prober in detector._mCharSetProbers:
+            if prober:
+                encodings[prober.get_charset_name()] = prober.get_confidence()
 
     return encodings
 
diff --git a/requirements.txt b/requirements.txt
index 217112b..551256f 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -6,7 +6,7 @@ binaryornot < 0.4.3
 bleach
 blinker
 celery
-chardet < 3.0.0
+chardet
 docutils
 enum34
 flask
diff --git a/tests/test_pagure_lib_encoding_utils.py b/tests/test_pagure_lib_encoding_utils.py
index 67fdb68..5c3d07d 100644
--- a/tests/test_pagure_lib_encoding_utils.py
+++ b/tests/test_pagure_lib_encoding_utils.py
@@ -26,13 +26,19 @@ class TestGuessEncoding(unittest.TestCase):
 
     def test_guess_encoding_favor_utf_8(self):
         """
-        Test that strings that could be UTF-8 or ISO-8859-2 result in UTF-8.
+        Test that strings that could be UTF-8 or ISO-8859-* result in UTF-8.
+
+        python-chardet-3.0.4-2.fc27.noarch detects it as ISO-8859-9
+        python-chardet-2.2.1-1.el7_1.noarch detects it as ISO-8859-2
         """
         data = u'Šabata'.encode('utf-8')
         result = encoding_utils.guess_encoding(data)
         chardet_result = chardet.detect(data)
         self.assertEqual(result, 'utf-8')
-        self.assertEqual(chardet_result['encoding'], 'ISO-8859-2')
+        if chardet.__version__[0] == '3':
+            self.assertEqual(chardet_result['encoding'], 'ISO-8859-9')
+        else:
+            self.assertEqual(chardet_result['encoding'], 'ISO-8859-2')
 
     def test_guess_encoding_no_data(self):
         """ Test encoding_utils.guess_encoding() with an empty string """
@@ -47,10 +53,20 @@ class TestGuessEncodings(unittest.TestCase):
         data = u'Šabata'.encode('utf-8')
         result = encoding_utils.guess_encodings(data)
         chardet_result = chardet.detect(data)
-        self.assertEqual(
-            [encoding.encoding for encoding in result],
-            ['utf-8', 'ISO-8859-2', 'windows-1252'])
-        self.assertEqual(chardet_result['encoding'], 'ISO-8859-2')
+        if chardet.__version__[0] == '3':
+            self.assertEqual(
+                [encoding.encoding for encoding in result],
+                ['utf-8', 'ISO-8859-9', 'ISO-8859-1', 'MacCyrillic',
+                 'IBM866', 'TIS-620', 'EUC-JP', 'EUC-KR', 'GB2312', 'KOI8-R',
+                 'Big5', 'IBM855', 'ISO-8859-7', 'SHIFT_JIS', 'windows-1253',
+                 'CP949', 'EUC-TW', 'ISO-8859-5', 'windows-1251',
+                 'windows-1255'])
+            self.assertEqual(chardet_result['encoding'], 'ISO-8859-9')
+        else:
+            self.assertEqual(
+                [encoding.encoding for encoding in result],
+                ['utf-8', 'ISO-8859-2', 'windows-1252'])
+            self.assertEqual(chardet_result['encoding'], 'ISO-8859-2')
 
     def test_guess_encodings_no_data(self):
         """ Test encoding_utils.guess_encodings() with an emtpy string """