diff --git a/emm/preprocessing/functions.py b/emm/preprocessing/functions.py
index b88c1c8..ed8cf91 100644
--- a/emm/preprocessing/functions.py
+++ b/emm/preprocessing/functions.py
@@ -19,11 +19,21 @@
 
 from __future__ import annotations
 
+import warnings
 from functools import partial
 from typing import Any, Callable
 
 import cleanco
-from unidecode import unidecode
+
+try:
+    from unidecode import unidecode
+except ImportError:
+    unidecode = None
+    warnings.warn(
+        "The 'unidecode' module is not installed. 'strip_accents_unicode' will default to an identity function. "
+        "Install 'unidecode' to enable accent stripping functionality.",
+        ImportWarning,
+    )
 
 from emm.preprocessing.abbreviation_util import abbreviations_to_words, legal_abbreviations_to_words
 
@@ -50,7 +60,7 @@ def map_shorthands(name):
 
     return {
         # Replace accented characters by their normalized representation, e.g. replace 'ä' with 'A\xa4'
-        "strip_accents_unicode": F.run_custom_function(unidecode),
+        "strip_accents_unicode": F.run_custom_function(unidecode if unidecode is not None else (lambda x: x)),
         # Replace all dash and underscore characters with a space characters
         "strip_hyphens": F.regex_replace(r"""[-_]""", " ", simple=True),
         # Replace all punctuation characters (e.g. '.', '-', '_', ''', ';') with spaces
diff --git a/pyproject.toml b/pyproject.toml
index 4ee6d89..89a8caa 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -31,8 +31,7 @@ dependencies = [
   "sparse-dot-topn>=1.1.1",
   "joblib",
   "pyarrow>=6.0.1", # seems to work with spark 3.1.2 - 3.3.1
-  "requests",
-  "unidecode"
+  "requests"
 ]
 dynamic = ["version"]
 
@@ -53,10 +52,14 @@ dev = [
   "pandoc",
   "pympler"
 ]
+preprocessing = [
+  "unidecode"
+]
 test = [
   "pytest",
   "pytest-ordering",
-  "virtualenv"
+  "virtualenv",
+  "unidecode"
 ]
 test-cov = [
   "coverage",