diff --git a/laser_encoders/laser_tokenizer.py b/laser_encoders/laser_tokenizer.py
index 0488cb2c..728fdde5 100644
--- a/laser_encoders/laser_tokenizer.py
+++ b/laser_encoders/laser_tokenizer.py
@@ -103,8 +103,8 @@ def tokenize_file(self, inp_fname: Path, out_fname: Path) -> None:
                     tokens = self.tokenize(line.strip())
                     file_out.write(tokens + "\n")
 
-    def __call__(self, text_or_batch, batch=False):
-        if not batch:
+    def __call__(self, text_or_batch):
+        if isinstance(text_or_batch, str):
             return self.tokenize(text_or_batch)
         else:
             return self.tokenize_batch(text_or_batch)
diff --git a/laser_encoders/models.py b/laser_encoders/models.py
index 037a4f9f..6d2a567f 100644
--- a/laser_encoders/models.py
+++ b/laser_encoders/models.py
@@ -94,10 +94,12 @@ def __init__(
         self.encoder.eval()
         self.sort_kind = sort_kind
 
-    def __call__(self, sentences):
+    def __call__(self, text_or_batch):
         if self.spm_model:
-            sentences = self.tokenizer(sentences)
-            return self.encode_sentences(sentences)
+            text_or_batch = self.tokenizer(text_or_batch)
+            if isinstance(text_or_batch, str):
+                text_or_batch = [text_or_batch]
+            return self.encode_sentences(text_or_batch)
         else:
             raise ValueError(
                 "Either initialize the encoder with an spm_model or pre-tokenize and use the encode_sentences method."
diff --git a/laser_encoders/test_laser_tokenizer.py b/laser_encoders/test_laser_tokenizer.py
index 1155f8d2..1350c108 100644
--- a/laser_encoders/test_laser_tokenizer.py
+++ b/laser_encoders/test_laser_tokenizer.py
@@ -65,6 +65,19 @@ def test_tokenize(tokenizer, input_text: str):
     assert tokenizer.tokenize(input_text) == expected_output
 
 
+def test_tokenizer_call_method(tokenizer, input_text: str):
+    single_string = "This is a test sentence."
+    expected_output = "▁this ▁is ▁a ▁test ▁sent ence ."
+    assert tokenizer(single_string) == expected_output
+
+    list_of_strings = ["This is a test sentence.", "This is another test sentence."]
+    expected_output = [
+        "▁this ▁is ▁a ▁test ▁sent ence .",
+        "▁this ▁is ▁another ▁test ▁sent ence .",
+    ]
+    assert tokenizer(list_of_strings) == expected_output
+
+
 def test_normalization(tokenizer):
     test_data = "Hello!!! How are you??? I'm doing great."
     expected_output = "▁hel lo !!! ▁how ▁are ▁you ??? ▁i ' m ▁do ing ▁great ."