diff --git a/.github/workflows/build.yaml b/.github/workflows/build.yaml
index 137dd29..8c4345d 100644
--- a/.github/workflows/build.yaml
+++ b/.github/workflows/build.yaml
@@ -23,3 +23,6 @@ jobs:
 
       - name: Install
         run: make install
+
+      - name: Test
+        run: make test
diff --git a/.gitignore b/.gitignore
index cd17d8a..ea5dc35 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,3 +1,4 @@
 venv
 *.egg-info
-build
\ No newline at end of file
+build
+__pycache__
\ No newline at end of file
diff --git a/Makefile b/Makefile
index 9fcce4d..57a6131 100644
--- a/Makefile
+++ b/Makefile
@@ -14,11 +14,14 @@ endif
 	python3 -m venv venv
 	$(VENV) pip install -r requirements.txt
 
+install-cli:
+	pip install .
+
+test:
+	python -m unittest discover -s tests
+
 transcribe: install
 	$(VENV) python transcribe_me/main.py
 
 transcribe-install: install
-	$(VENV) python transcribe_me/main.py install
-
-install-cli:
-	pip install .
\ No newline at end of file
+	$(VENV) python transcribe_me/main.py install
\ No newline at end of file
diff --git a/tests/__init__.py b/tests/__init__.py
new file mode 100644
index 0000000..e69de29
diff --git a/tests/test_main.py b/tests/test_main.py
new file mode 100644
index 0000000..c1742ed
--- /dev/null
+++ b/tests/test_main.py
@@ -0,0 +1,101 @@
+import unittest
+from unittest.mock import patch, mock_open
+import argparse
+from transcribe_me import main
+
+class MockAudioSegment:
+    def __init__(self, length):
+        self.length = length
+
+    def __len__(self):
+        return self.length
+
+    def __getitem__(self, index):
+        return self
+
+    def export(self, *args, **kwargs):
+        pass
+
+class MockTranscription:
+    def __init__(self, text):
+        self.text = text
+
+class MockAnthropic:
+    class Messages:
+        def create(self, *args, **kwargs):
+            return 'summary'
+    messages = Messages()
+
+class TestMain(unittest.TestCase):
+
+    @patch('os.path.splitext')
+    @patch('transcribe_me.main.AudioSegment.from_file')
+    def test_split_audio(self, mock_from_file, mock_splitext):
+        mock_splitext.return_value = ('file', '.mp3')
+        mock_from_file.return_value = MockAudioSegment(1000)
+        result = main.split_audio('file.mp3')
+        self.assertEqual(result, ['file_part1.mp3'])
+
+    @patch('openai.audio.transcriptions.create')
+    @patch('builtins.open', new_callable=mock_open, read_data="data")
+    def test_transcribe_chunk(self, mock_open, mock_create):
+        mock_transcription = MockTranscription('transcription')
+        mock_create.return_value = mock_transcription
+        result = main.transcribe_chunk('file.mp3')
+        self.assertEqual(result, 'transcription')
+
+    @patch('os.remove')
+    @patch('transcribe_me.main.split_audio')
+    @patch('transcribe_me.main.transcribe_chunk')
+    @patch('builtins.open', new_callable=mock_open)
+    def test_transcribe_audio(self, mock_open, mock_transcribe_chunk, mock_split_audio, mock_remove):
+        mock_split_audio.return_value = ['file_part1.mp3']
+        mock_transcription = MockTranscription('transcription')
+        mock_transcribe_chunk.return_value = mock_transcription
+        main.transcribe_audio('file.mp3', 'output.txt')
+        mock_open.assert_called_with('output.txt', 'w', encoding='utf-8')
+        mock_remove.assert_called_with('file_part1.mp3')
+
+    # @patch('transcribe_me.main.anthropic.Anthropic', new=MockAnthropic())
+    # def test_generate_summary_openai(self):
+    #     model_config = {
+    #         "temperature": 0.1,
+    #         "max_tokens": 2048,
+    #         "model": "gpt-4",
+    #         "system_prompt": "Generate a summary with key points in bold and a Next Steps section, use Markdown, be a concise tech expert but kind to non-technical readers.",
+    #     }
+    #     result = main.generate_summary('transcription', model_config)
+    #     self.assertEqual(result, 'summary')
+
+    # @patch('anthropic.Anthropic')
+    # def test_generate_summary_anthropic(self, mock_anthropic):
+    #     mock_anthropic.return_value = 'summary'
+    #     model_config = {
+    #         "temperature": 0.1,
+    #         "max_tokens": 2048,
+    #         "model": "claude-3-sonnet-20240229",
+    #         "system_prompt": "Generate a summary with key points in bold and a Next Steps section, use Markdown, be a concise tech expert but kind to non-technical readers.",
+    #     }
+    #     result = main.generate_summary('transcription', model_config)
+    #     self.assertEqual(result, 'summary')
+
+    # @patch('argparse.ArgumentParser.parse_args')
+    # @patch('os.makedirs')
+    # @patch('os.listdir')
+    # @patch('os.path.exists')
+    # @patch('transcribe_me.main.transcribe_audio')
+    # @patch('transcribe_me.main.generate_summary')
+    # @patch('builtins.open', new_callable=mock_open, read_data="data")
+    # def test_main(self, mock_open, mock_generate_summary, mock_transcribe_audio, mock_exists, mock_listdir, mock_makedirs, mock_parse_args):
+    #     mock_parse_args.return_value = argparse.Namespace(command=None, input='input', output='output')
+    #     mock_listdir.return_value = ['file1', 'file2']
+    #     mock_exists.side_effect = [False, False, True, True]
+    #     mock_generate_summary.return_value = 'summary'
+    #     main.main()
+    #     mock_makedirs.assert_called()
+    #     mock_transcribe_audio.assert_called()
+    #     mock_generate_summary.assert_called()
+    #     mock_open.assert_called()
+
+if __name__ == '__main__':
+    unittest.main()
\ No newline at end of file
diff --git a/transcribe_me/main.py b/transcribe_me/main.py
index 55446cb..9f1babd 100644
--- a/transcribe_me/main.py
+++ b/transcribe_me/main.py
@@ -1,8 +1,7 @@
 import argparse
 import os
 from glob import glob
-from typing import Tuple, Dict, Any
-import sys
+from typing import Dict, Any
 
 import anthropic
 import openai