diff --git a/CHANGELOG.md b/CHANGELOG.md
index 65d44f4..b1a56dc 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -5,6 +5,9 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/)
 and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.html).
 
 ## [Unreleased]
+### Added
+- `TocountError` class
+- `TocountValidationError` class
 ### Changed
 - `setup.py` updated
 ## [0.5] - 2026-01-02
diff --git a/tests/test_rule_based.py b/tests/test_rule_based.py
index 93fa2f7..70f9d82 100644
--- a/tests/test_rule_based.py
+++ b/tests/test_rule_based.py
@@ -1,5 +1,6 @@
 import pytest
 from tocount import estimate_text_tokens, TextEstimator
+from tocount import TocountValidationError
 from tocount.params import INVALID_TEXT_MESSAGE, INVALID_TEXT_ESTIMATOR_MESSAGE
 
 
@@ -135,12 +136,12 @@ def test_openai_gpt_4_text_with_non_english():
 
 def test_raises_error_for_invalid_text():
     invalid_text = 12345
-    with pytest.raises(ValueError, match=INVALID_TEXT_MESSAGE):
+    with pytest.raises(TocountValidationError, match=INVALID_TEXT_MESSAGE):
         estimate_text_tokens(invalid_text)
 
 
 def test_raises_error_for_invalid_estimator():
     valid_text = "sample prompt"
     invalid_estimator = "not a valid estimator"
-    with pytest.raises(ValueError, match=INVALID_TEXT_ESTIMATOR_MESSAGE):
+    with pytest.raises(TocountValidationError, match=INVALID_TEXT_ESTIMATOR_MESSAGE):
         estimate_text_tokens(valid_text, invalid_estimator)
diff --git a/tocount/__init__.py b/tocount/__init__.py
index 50ef7d4..70382cd 100644
--- a/tocount/__init__.py
+++ b/tocount/__init__.py
@@ -2,7 +2,8 @@
 """Tocount modules."""
 
 from .params import TOCOUNT_VERSION, TextEstimator
+from .errors import TocountError, TocountValidationError
 from .functions import estimate_text_tokens
 __version__ = TOCOUNT_VERSION
 
-__all__ = ["TextEstimator", "estimate_text_tokens"]
+__all__ = ["TextEstimator", "estimate_text_tokens", "TocountError", "TocountValidationError"]
diff --git a/tocount/errors.py b/tocount/errors.py
new file mode 100644
index 0000000..d6029f2
--- /dev/null
+++ b/tocount/errors.py
@@ -0,0 +1,12 @@
+# -*- coding: utf-8 -*-
+"""Tocount errors."""
+
+class TocountError(Exception):
+    """Base exception for all Tocount errors."""
+
+    pass
+
+class TocountValidationError(TocountError, ValueError):
+    """Base class for validation errors in Tocount."""
+
+    pass
diff --git a/tocount/functions.py b/tocount/functions.py
index 3f68487..bb69dcf 100644
--- a/tocount/functions.py
+++ b/tocount/functions.py
@@ -4,6 +4,7 @@
 from .params import TextEstimator, _TextEstimatorRuleBased, _TextEstimatorTikTokenR50K
 from .params import _TextEstimatorTikTokenCL100K, _TextEstimatorTikTokenO200K
 from .params import _TextEstimatorDeepseekR1, _TextEstimatorQwenQwQ, _TextEstimatorLlama_3_1
+from .errors import TocountValidationError
 from .rule_based.functions import universal_tokens_estimator, openai_tokens_estimator_gpt_3_5, openai_tokens_estimator_gpt_4
 from .tiktoken_r50k.functions import linear_tokens_estimator_all as r50k_linear_all
 from .tiktoken_r50k.functions import linear_tokens_estimator_english as r50k_linear_english
@@ -47,7 +48,7 @@ def estimate_text_tokens(text: str, estimator: TextEstimator = TextEstimator.DEF
     :return: tokens number
     """
     if not isinstance(text, str):
-        raise ValueError(INVALID_TEXT_MESSAGE)
+        raise TocountValidationError(INVALID_TEXT_MESSAGE)
     if not isinstance(estimator, (TextEstimator, _TextEstimatorRuleBased, _TextEstimatorTikTokenR50K, _TextEstimatorTikTokenCL100K, _TextEstimatorTikTokenO200K, _TextEstimatorDeepseekR1, _TextEstimatorQwenQwQ, _TextEstimatorLlama_3_1)):
-        raise ValueError(INVALID_TEXT_ESTIMATOR_MESSAGE)
+        raise TocountValidationError(INVALID_TEXT_ESTIMATOR_MESSAGE)
     return text_estimator_map[estimator](text)