diff --git a/build/lib/markovbot/__init__.py b/build/lib/markovbot/__init__.py
deleted file mode 100644
index d33a348..0000000
--- a/build/lib/markovbot/__init__.py
+++ /dev/null
@@ -1,29 +0,0 @@
-import argparse
-
-def parser() =
-    parser = argparse.ArgumentParser(description='Process some integers.')
-    parser.add_argument('--config', dest='filepath', metavar='CREDENTIALS', type=str, help='Path to config file')
-    parser.add_argument('text', metavar='CORPUS', type=str, help='path to the text you wish to mimic') # filepath should be a flag; corpus an argument!
-    parser.add_argument('--coke-binge', dest='coke_binge', action="store_true", default=False, help='Tweet excessively/emulate our POTUS') # instead of tweeting once, tweet a random number of times.
-    parser.add_argument('-t', dest='test', action="store_true", default=False, help='Test text generation without tweeting.') # distance on markov chains?? distance on syntax trees? # default to test or default to tweet??
-    return parser.parse_args()
-
-# Function to help process credentials file
-def drop_tag(str):
-    parsed = itertools.dropwhile(lambda x: x!= ':', str)
-    str2 = "".join(parsed)
-    parsed2 = itertools.dropwhile(lambda x: x == ' ' or x == ':', str2)
-    return "".join(parsed2)
-
-# Process api keys etc.
-def make_api_keys(filepath):
-    # Open file with credentials
-    with open(filepath) as f:
-        content = f.readlines()
-    # Strip irrelevant text
-    let content = [drop_tag(x.strip()) for x in content]
-    # create twitter api object
-    return twitter.Api(consumer_key=content[0],
-                          consumer_secret=content[1],
-                          access_token_key=content[2],
-                          access_token_secret=content[3])
diff --git a/build/scripts-3.5/markovbot b/build/scripts-3.5/markovbot
deleted file mode 100755
index d31776c..0000000
--- a/build/scripts-3.5/markovbot
+++ /dev/null
@@ -1,41 +0,0 @@
-#!python
-import markovify
-import markovbot
-import twitter
-import itertools
-import random
-
-# Command line parser
-
-args = parser()
-
-# Set the coke binge parameter
-if args.coke_binge:
-    coke_binge_num = random.randint(1, 6)
-else:
-    coke_binge_num = 1
-
-# Get raw text of markov chain as string.
-with open(args.text) as f:
-    text = f.read()
-
-# Build the model.
-text_model = markovify.Text(text)
-
-api = make_api_keys(args.filepath)
-
-# Generate three randomly-generated sentences of no more than 140 characters and tweet them.
-selected = False
-for i in range(0, coke_binge_num):
-    while selected is False:
-        str_potentially = text_model.make_short_sentence(140)
-        if not("http" in str_potentially):
-            if not(args.test):
-                # tweet the generated text
-                status = api.PostUpdate(str_potentially)
-                print(status.text)  # verify it worked
-            else:
-                # test mode; just display it
-                print(str_potentially)
-            selected = True
-    selected = False
diff --git a/setup.py b/setup.py
index 2923138..2fd4bca 100755
--- a/setup.py
+++ b/setup.py
@@ -10,7 +10,6 @@ setup(name='markovbot',
       packages=['markovbot'],
       scripts=['bin/markovbot'],
       install_requires=[
-          'gitpython',
           'markovify',
           'python-twitter',
       ],