diff --git a/README.md b/README.md
index a72f6da..59ef71d 100644
--- a/README.md
+++ b/README.md
@@ -2,6 +2,15 @@
 
 Codegreen is a Python package and CLI tool for energy profiling of Python code. It allows measuring the energy consumption of Deep learning framework APIs like TensorFlow.
 
+## Prerequisites
+
+Codegreen requires the following hardware:
+
+- Intel processor
+- NVIDIA GPU 
+
+The Intel processor is needed for accessing the RAPL power sensors. An NVIDIA GPU is required for profiling GPU usage via `nvidia-smi`.
+
 ## Installation
 
 ```
@@ -47,12 +56,24 @@ Usage: codegreen [OPTIONS] COMMAND [ARGS]...
 
 Basic usage:
 
+```bash
+codegreen project-patcher --project myscripts/
 ```
-codegreen project-patcher myscripts/
+
+Patch all Python scripts in `myscripts/` for instrumentation. Patched projects are saved with a `_patched` suffix.
+
+```bash  
 codegreen run-energy-profiler
-codegreen start-energy-measurement
 ```
 
+This will start the energy measurement server.
+
+```bash
+codegreen start-energy-measurement --project repo_dir --scripts script_to_run --iterations "5"
+``` 
+
+This will run the patched scripts for project `repo_dir`, scripts `script_to_run` and number of run iterations `n` and collect power usage data. Save the energy profiles and execution logs under `experiments/`.
+
 ### As a Python package
 
 ```python
@@ -71,9 +92,10 @@ See the API documentation for more details.
     - `patching/`: Code instrumentation
     - `measurement/`: Measurement and output
     - `experiment/`: Experimental configurations
-- `dist/`: Built distributions
+  - `utils/`: utility functions
+<!-- - `dist/`: Built distributions
 - `tests/`: Unit tests
-- `examples/`: Usage examples
+- `examples/`: Usage examples -->
 
 <!-- 
 ## Contributing
diff --git a/test/test_main.py b/test/test_main.py
deleted file mode 100644
index 0589c69..0000000
--- a/test/test_main.py
+++ /dev/null
@@ -1,63 +0,0 @@
-import os
-import shutil
-import subprocess
-from pathlib import Path
-from unittest import mock
-
-import pytest
-import typer
-
-from codegreen.main import start_energy_measurement
-
-
-@pytest.fixture
-def project_dir(tmp_path):
-    # Create a temporary directory for the project
-    project_dir = tmp_path / "my_project"
-    project_dir.mkdir()
-
-    # Create a sample script
-    script_path = project_dir / "script.py"
-    script_path.write_text("print('Hello, world!')")
-
-    return project_dir
-
-
-def test_start_energy_measurement(project_dir):
-    # Define the arguments for the function
-    project = project_dir
-    scripts = [project_dir / "script.py"]
-
-    # Mock the get_repo_metadata function to return a dummy value
-    with mock.patch("codegreen.main.get_repo_metadata") as mock_get_repo_metadata:
-        mock_get_repo_metadata.return_value = {"author": "John Doe"}
-
-        # Call the function
-        start_energy_measurement(project=project, scripts=scripts)
-
-    # Check that the patched files were created
-    patched_dir = project_dir / "my_project_patched"
-    assert patched_dir.exists()
-
-    # Check that the method-level scripts were created
-    method_level_scripts_dir = patched_dir / "method-level-scripts"
-    assert method_level_scripts_dir.exists()
-
-    # Check that the project-level scripts were created
-    project_level_scripts_dir = patched_dir / "project-level-scripts"
-    assert project_level_scripts_dir.exists()
-
-    # Check that the original scripts were not modified
-    assert (project_dir / "script.py").read_text() == "print('Hello, world!')"
-
-    # Check that the method-level script was run successfully
-    output = subprocess.check_output(["python3", str(method_level_scripts_dir / "script_method-level.py"), "1", str(project_dir / "script")])
-    assert output == b"Hello, world!\n"
-
-    # Check that the log file was created
-    log_file = patched_dir / "codegreen.log"
-    assert log_file.exists()
-
-    # Check that the patched directory was deleted
-    shutil.rmtree(patched_dir)
-    assert not patched_dir.exists()
\ No newline at end of file