diff --git a/allennlp/commands/__init__.py b/allennlp/commands/__init__.py
index 92ce75e9782..7922b3fe31c 100644
--- a/allennlp/commands/__init__.py
+++ b/allennlp/commands/__init__.py
@@ -1,7 +1,6 @@
 from typing import Dict
 import argparse
 import logging
-import sys
 
 from allennlp.commands.elmo import Elmo
 from allennlp.commands.evaluate import Evaluate
@@ -11,8 +10,8 @@
 from allennlp.commands.serve import Serve
 from allennlp.commands.dry_run import DryRun
 from allennlp.commands.subcommand import Subcommand
+from allennlp.commands.test_install import TestInstall
 from allennlp.commands.train import Train
-from allennlp.service.predictors import DemoModel
 from allennlp.common.util import import_submodules
 
 logger = logging.getLogger(__name__)  # pylint: disable=invalid-name
@@ -40,6 +39,7 @@ def main(prog: str = None,
             "elmo": Elmo(),
             "fine-tune": FineTune(),
             "dry-run": DryRun(),
+            "test-install": TestInstall(),
 
             # Superseded by overrides
             **subcommand_overrides
diff --git a/allennlp/commands/test_install.py b/allennlp/commands/test_install.py
new file mode 100644
index 00000000000..03421d49ef5
--- /dev/null
+++ b/allennlp/commands/test_install.py
@@ -0,0 +1,66 @@
+"""
+The ``test-install`` subcommand verifies
+an installation by running the unit tests.
+
+.. code-block:: bash
+
+    $ allennlp test-install --help
+    usage: allennlp test-install [-h] [--run-all]
+                                 [--include-package INCLUDE_PACKAGE]
+
+    Test that installation works by running the unit tests.
+
+    optional arguments:
+      -h, --help            show this help message and exit
+      --run-all             By default, we skip tests that are slow or download
+                            large files. This flag will run all tests.
+      --include-package INCLUDE_PACKAGE
+                            additional packages to include
+"""
+
+import argparse
+import logging
+import os
+
+import pytest
+
+from allennlp.commands.subcommand import Subcommand
+
+logger = logging.getLogger(__name__)  # pylint: disable=invalid-name
+
+class TestInstall(Subcommand):
+    def add_subparser(self, name: str, parser: argparse._SubParsersAction) -> argparse.ArgumentParser:
+        # pylint: disable=protected-access
+        description = '''Test that installation works by running the unit tests.'''
+        subparser = parser.add_parser(
+                name, description=description, help='Run the unit tests.')
+
+        subparser.add_argument('--run-all', action="store_true",
+                               help="By default, we skip tests that are slow "
+                               "or download large files. This flag will run all tests.")
+
+        subparser.set_defaults(func=_run_test)
+
+        return subparser
+
+
+def _get_project_root():
+    return os.path.abspath(
+            os.path.join(
+                    os.path.dirname(os.path.realpath(__file__)),
+                    os.pardir, os.pardir))
+
+
+def _run_test(args: argparse.Namespace):
+    initial_working_dir = os.getcwd()
+    project_root = _get_project_root()
+    logger.info("Changing directory to %s", project_root)
+    os.chdir(project_root)
+    test_dir = os.path.join(project_root, "tests")
+    logger.info("Running tests at %s", test_dir)
+    if args.run_all:
+        pytest.main([test_dir])
+    else:
+        pytest.main([test_dir, '-k', 'not sniff_test and not notebooks_test'])
+    # Change back to original working directory after running tests
+    os.chdir(initial_working_dir)
diff --git a/doc/api/allennlp.commands.rst b/doc/api/allennlp.commands.rst
index 2761bbecdd4..865c97ac806 100644
--- a/doc/api/allennlp.commands.rst
+++ b/doc/api/allennlp.commands.rst
@@ -17,14 +17,17 @@ The included module ``allennlp.run`` is such a script:
       -h, --help  show this help message and exit
 
     Commands:
-        train     Train a model
-        evaluate  Evaluate the specified model + dataset
-        predict   Use a trained model to make predictions.
-        serve     Run the web service and demo.
-        make-vocab
-                  Create a vocabulary
-        elmo      Use a trained model to make predictions.
-        dry-run   Create a vocabulary, compute dataset statistics and other training utilities.
+        train       Train a model
+        evaluate    Evaluate the specified model + dataset
+        predict     Use a trained model to make predictions.
+        serve       Run the web service and demo.
+        make-vocab  Create a vocabulary
+        elmo        Use a trained model to make predictions.
+        fine-tune   Continue training a model on a new dataset
+        dry-run     Create a vocabulary, compute dataset statistics and other
+                    training utilities.
+        test-install
+                    Run the unit tests.
 
 However, it only knows about the models and classes that are
 included with AllenNLP. Once you start creating custom models,
@@ -41,6 +44,7 @@ calls ``main()``.
     allennlp.commands.fine_tune
     allennlp.commands.elmo
     allennlp.commands.dry_run
+    allennlp.commands.test_install
 
 .. automodule:: allennlp.commands
    :members:
diff --git a/doc/api/allennlp.commands.test_install.rst b/doc/api/allennlp.commands.test_install.rst
new file mode 100644
index 00000000000..40fda6b8687
--- /dev/null
+++ b/doc/api/allennlp.commands.test_install.rst
@@ -0,0 +1,4 @@
+allennlp.commands.test_install
+==============================
+
+.. automodule:: allennlp.commands.test_install
diff --git a/requirements.txt b/requirements.txt
index b0cc945fe0a..8df25db27c7 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -66,3 +66,14 @@ h5py
 
 # For timezone utilities
 pytz==2017.3
+
+#### ESSENTIAL TESTING-RELATED PACKAGES ####
+# We'll use pytest to run our tests; this isn't really necessary to run the code, but it is to run
+# the tests.  With this here, you can run the tests with `py.test` from the base directory.
+pytest
+
+# Allows marking tests as flaky, to be rerun if they fail
+flaky
+
+# Required to mock out `requests` calls
+responses>=0.7
diff --git a/requirements_test.txt b/requirements_test.txt
index 3905175b992..403f6be02ed 100644
--- a/requirements_test.txt
+++ b/requirements_test.txt
@@ -3,10 +3,6 @@
 # Checks style, syntax, and other useful errors.
 pylint==1.8.1
 
-# We'll use pytest to run our tests; this isn't really necessary to run the code, but it is to run
-# the tests.  With this here, you can run the tests with `py.test` from the base directory.
-pytest
-
 # Tutorial notebooks
 jupyter
 
@@ -16,9 +12,6 @@ mypy==0.521
 # Allows generation of coverage reports with pytest.
 pytest-cov
 
-# Allows marking tests as flaky, to be rerun if they fail
-flaky
-
 # Allows codecov to generate coverage reports
 coverage
 codecov
@@ -26,9 +19,6 @@ codecov
 # Required to run sanic tests
 aiohttp
 
-# Required to mock out `requests` calls
-responses>=0.7
-
 #### DOC-RELATED PACKAGES ####
 
 # Builds our documentation.
diff --git a/setup.py b/setup.py
index 8335eed245f..7341ecac9c0 100644
--- a/setup.py
+++ b/setup.py
@@ -120,7 +120,10 @@
           'scikit-learn',
           'scipy',
           'pytz==2017.3',
-          'unidecode'
+          'unidecode',
+          'pytest',
+          'flaky',
+          'responses>=0.7'
       ],
       scripts=["bin/allennlp"],
       setup_requires=setup_requirements,
diff --git a/tests/commands/test_install_test.py b/tests/commands/test_install_test.py
new file mode 100644
index 00000000000..0fb56546a8a
--- /dev/null
+++ b/tests/commands/test_install_test.py
@@ -0,0 +1,13 @@
+# pylint: disable=invalid-name,no-self-use
+import os
+
+from allennlp.common.testing import AllenNlpTestCase
+from allennlp.commands.test_install import _get_project_root
+
+
+class TestTestInstall(AllenNlpTestCase):
+    def test_get_project_root(self):
+        project_root = _get_project_root()
+        assert os.path.exists(os.path.join(project_root, "tests"))
+        assert os.path.exists(os.path.join(project_root, "LICENSE"))
+        assert os.path.exists(os.path.join(project_root, "setup.py"))
diff --git a/tests/notebooks_test.py b/tests/notebooks_test.py
index 9d23a76439e..4d999b5dc19 100644
--- a/tests/notebooks_test.py
+++ b/tests/notebooks_test.py
@@ -1,8 +1,12 @@
 import os
 
-import nbformat
-from nbconvert.preprocessors.execute import CellExecutionError
-from nbconvert.preprocessors import ExecutePreprocessor
+try:
+    import nbformat
+    from nbconvert.preprocessors.execute import CellExecutionError
+    from nbconvert.preprocessors import ExecutePreprocessor
+except ModuleNotFoundError:
+    print("jupyter must be installed in order to run notebook tests. "
+          "To install with pip, run: pip install jupyter")
 
 from allennlp.common.testing import AllenNlpTestCase