diff --git a/MANIFEST.in b/MANIFEST.in
index 1641225..6c6bdd6 100644
--- a/MANIFEST.in
+++ b/MANIFEST.in
@@ -1,3 +1,6 @@
-include Makefile
+include Makefile gittaggers.py
+include tests/*
+include tests/test-data/*
 global-exclude *~
 global-exclude *.pyc
+
diff --git a/Makefile b/Makefile
index f377b2f..044b2e6 100644
--- a/Makefile
+++ b/Makefile
@@ -26,7 +26,7 @@ MODULE=cwltest
 # `SHELL=bash` doesn't work for some, so don't use BASH-isms like
 # `[[` conditional expressions.
 PYSOURCES=$(wildcard ${MODULE}/**.py tests/*.py) setup.py
-DEVPKGS=pep8 diff_cover autopep8 pylint coverage pep257 flake8
+DEVPKGS=pep8 diff_cover autopep8 pylint coverage pep257 flake8 pytest
 DEBDEVPKGS=pep8 python-autopep8 pylint python-coverage pep257 sloccount python-flake8
 VERSION=1.0.$(shell date +%Y%m%d%H%M%S --date=`git log --first-parent \
 	--max-count=1 --format=format:%cI`)
@@ -34,7 +34,7 @@ mkfile_dir := $(dir $(abspath $(lastword $(MAKEFILE_LIST))))
 
 ## all         : default task
 all:
-	./setup.py develop
+	pip install -e .
 
 ## help        : print this help message and exit
 help: Makefile
@@ -109,10 +109,7 @@ diff_pylint_report: pylint_report.txt
 	diff-quality --violations=pylint pylint_report.txt
 
 .coverage: $(PYSOURCES) all
-	export COVERAGE_PROCESS_START=${mkfile_dir}.coveragerc; \
-	       cd ${CWL}; ./run_test.sh RUNNER=cwltool
-	coverage run setup.py test
-	coverage combine ${CWL} ${CWL}/draft-3/ ./
+	coverage run ./setup.py test
 
 coverage.xml: .coverage
 	python-coverage xml
@@ -135,6 +132,7 @@ diff-cover.html: coverage-gcovr.xml coverage.xml
 
 ## test        : run the ${MODULE} test suite
 test: FORCE
+	./setup.py test
 
 sloccount.sc: ${PYSOURCES} Makefile
 	sloccount --duplicates --wide --details $^ > sloccount.sc
diff --git a/release-test.sh b/release-test.sh
index c0c4ff2..124a2ea 100755
--- a/release-test.sh
+++ b/release-test.sh
@@ -6,7 +6,7 @@ set -x
 package=cwltest
 module=cwltest
 repo=https://github.com/common-workflow-language/cwltest.git
-run_tests="cwltest --help"
+run_tests="py.test --pyarg ${module}"
 pipver=8.0.1 # minimum required version of pip
 setupver=20.10.1 # minimum required version of setuptools
 
diff --git a/setup.cfg b/setup.cfg
index 87c34d2..903a610 100644
--- a/setup.cfg
+++ b/setup.cfg
@@ -5,3 +5,9 @@ ignore = E124,E128,E129,E201,E202,E225,E226,E231,E265,E271,E302,E303,F401,E402,E
 
 [bdist_wheel]
 universal = 1
+
+[tool:pytest]
+testpaths = tests
+
+[aliases]
+test=pytest
diff --git a/setup.py b/setup.py
index 9127eb0..0c05388 100755
--- a/setup.py
+++ b/setup.py
@@ -22,6 +22,9 @@
     'six>=1.10.0'
 ]
 
+needs_pytest = {'pytest', 'test', 'ptr'}.intersection(sys.argv)
+pytest_runner = ['pytest-runner'] if needs_pytest else []
+
 if sys.version_info.major == 2:
     install_requires.extend(['futures >= 3.0.5', 'subprocess32'])
 
@@ -37,10 +40,13 @@
       url="https://github.com/common-workflow-language/cwltest",
       download_url="https://github.com/common-workflow-language/cwltest",
       license='Apache 2.0',
-      packages=["cwltest"],
+      packages=["cwltest", "cwltest.tests"],
+      package_dir={'cwltest.tests': 'tests'},
+      include_package_data=True,
       install_requires=install_requires,
       test_suite='tests',
-      tests_require=[],
+      setup_requires=[] + pytest_runner,
+      tests_require=['pytest'],
       entry_points={
           'console_scripts': ["cwltest=cwltest:main"]
       },
diff --git a/tests/test_compare.py b/tests/test_compare.py
index ddaeeea..d748430 100644
--- a/tests/test_compare.py
+++ b/tests/test_compare.py
@@ -1,18 +1,10 @@
 import unittest
+from .util import get_data
 from cwltest import CompareFail
 from cwltest.utils import compare_file, compare
 
 
-class TestCompare(unittest.TestCase):
-
-    def compare_success(self, expected, actual):
-        try:
-            compare(expected, actual)
-        except CompareFail:
-            self.fail("Comparison failed unexpectedly")
-
-
-class TestCompareFile(TestCompare):
+class TestCompareFile(unittest.TestCase):
 
     def test_compare_file(self):
         expected = {
@@ -20,7 +12,7 @@ def test_compare_file(self):
             "size": 2,
             "class": "File",
             "checksum": "sha1$7448d8798a4380162d4b56f9b452e2f6f9e24e7a"
-         }
+        }
 
         actual = {
             "basename": "cores.txt",
@@ -48,10 +40,10 @@ def test_compare_contents_success(self):
             "checksum": "sha1$7448d8798a4380162d4b56f9b452e2f6f9e24e7a",
             "class": "File",
             "location": "file:///var/folders/8x/2df05_7j20j6r8y81w4qf43r0000gn/T/tmpG0EkrS/cores.txt",
-            "path": "tests/test-data/cores.txt",
+            "path": get_data("tests/test-data/cores.txt"),
             "size": 2
         }
-        self.compare_success(expected, actual)
+        compare(expected, actual)
 
     def test_compare_contents_failure(self):
         expected = {
@@ -66,7 +58,7 @@ def test_compare_contents_failure(self):
             "checksum": "sha1$7448d8798a4380162d4b56f9b452e2f6f9e24e7a",
             "class": "File",
             "location": "file:///var/folders/8x/2df05_7j20j6r8y81w4qf43r0000gn/T/tmpG0EkrS/cores.txt",
-            "path": "tests/test-data/cores.txt",
+            "path": get_data("tests/test-data/cores.txt"),
             "size": 2
         }
         with self.assertRaises(CompareFail):
diff --git a/tests/util.py b/tests/util.py
new file mode 100644
index 0000000..c131cd3
--- /dev/null
+++ b/tests/util.py
@@ -0,0 +1,19 @@
+from __future__ import absolute_import
+import os
+
+from pkg_resources import (Requirement, ResolutionError,  # type: ignore
+                           resource_filename)
+
+
+def get_data(filename):
+    filename = os.path.normpath(
+        filename)  # normalizing path depending on OS or else it will cause problem when joining path
+    filepath = None
+    try:
+        filepath = resource_filename(
+            Requirement.parse("cwltest"), filename)
+    except ResolutionError:
+        pass
+    if not filepath or not os.path.isfile(filepath):
+        filepath = os.path.join(os.path.dirname(__file__), os.pardir, filename)
+    return filepath
diff --git a/tox.ini b/tox.ini
index 348a5f3..7af39d7 100644
--- a/tox.ini
+++ b/tox.ini
@@ -1,6 +1,6 @@
 [tox]
 #envlist = py35-lint,py34-lint,py33-lint,py27-lint,py35-unit,py34-unit,py33-unit,py27-unit
-envlist = py27-lint, py27-unit, py35-mypy{2,3}
+envlist = py27-lint, py{27,33,34,35}-unit, py35-mypy{2,3}
 skipsdist = True
 
 [travis]
@@ -12,7 +12,9 @@ python =
   3.5: py35-mypy{2,3}
 
 [testenv]
-deps = -rrequirements.txt
+deps =
+  pytest
+  -rrequirements.txt
 
 [testenv:py35-mypy2]
 commands = make mypy2
@@ -49,13 +51,14 @@ whitelist_externals = flake8
 deps = flake8
 
 [testenv:py35-unit]
-commands = python setup.py test
+commands = pytest {posargs} # substitute with tox' positional arguments
 
 [testenv:py34-unit]
-commands = python setup.py test
+commands = pytest {posargs} # substitute with tox' positional arguments
 
 [testenv:py33-unit]
-commands = python setup.py test
+commands = pytest {posargs} # substitute with tox' positional arguments
 
 [testenv:py27-unit]
-commands = python setup.py test
+commands = pytest {posargs} # substitute with tox' positional arguments
+