diff --git a/Pickle/freq.pkl b/Pickle/freq.pkl
new file mode 100644
index 0000000..266a419
Binary files /dev/null and b/Pickle/freq.pkl differ
diff --git a/Pickle/rules.pkl b/Pickle/rules.pkl
new file mode 100644
index 0000000..49e3521
Binary files /dev/null and b/Pickle/rules.pkl differ
diff --git a/Rest_API/server.py b/Rest_API/server.py
index 3d7b6f9..0876073 100644
--- a/Rest_API/server.py
+++ b/Rest_API/server.py
@@ -2,21 +2,26 @@
 import pickle
 
 app = Flask(__name__)
-
-rules_path = 'rules.pkl'
+rules_path = '../Pickle/rules.pkl'
+freq_path = '../Pickle/freq.pkl'
 
 @app.route('/api/recommend', methods=['POST'])
 def recommend():
     songs = request.json['songs']
-
     rules = pickle.load(open(rules_path, 'rb'))
+    freq = pickle.load(open(freq_path, 'rb'))
+    recommended_playlists = []
 
-    recommended_songs = []
-
+    # Verifica as regras de associação para recomendar playlists
     for song in songs:
-        recommended_songs.append("teste")
+        filtered_rules = rules[rules['antecedents'] == frozenset({song})]
+        if not filtered_rules.empty:
+            recommended_playlists.extend(filtered_rules['consequents'].values[0])
+
+    # Remove duplicatas e converte para uma lista única de playlists recomendadas
+    recommended_playlists = list(set(recommended_playlists))
 
-    return jsonify(recommended_songs)
+    return jsonify({'recommended_playlists': recommended_playlists})
 
 if __name__ == '__main__':
     app.run(port=32196)
\ No newline at end of file
diff --git a/Train_model/ml.py b/Train_model/ml.py
index 9f8c1a1..9ce93a7 100644
--- a/Train_model/ml.py
+++ b/Train_model/ml.py
@@ -32,19 +32,17 @@ def encode(self, x):
     
     def get_freq(self):
         self.freq = apriori(self.df, min_support=min_support, use_colnames=True)
-        self.freq.to_pickle('freq.pkl')
+        self.freq.to_pickle('../Pickle/freq.pkl')
         print("freq.pkl created successfully")
     
     def get_rules(self):
         self.rules = association_rules(self.freq, metric="lift", min_threshold=min_threshold)
-        self.rules.to_pickle('rules.pkl')
+        self.rules.to_pickle('../Pickle/rules.pkl')
         print("rules.pkl created successfully")
 
 
 if __name__ == '__main__':
     fdm = FreqDatasetMining(FILE_PATH)
-
     fdm.group_by_pid_and_track_uri()
-
     fdm.get_freq()
     fdm.get_rules()
\ No newline at end of file
diff --git a/Rest_API/client.py b/client.py
similarity index 77%
rename from Rest_API/client.py
rename to client.py
index be3c380..d2b8aa8 100644
--- a/Rest_API/client.py
+++ b/client.py
@@ -4,6 +4,10 @@
 
 api_url = 'http://127.0.0.1:32196/api/recommend'
 
+if len(sys.argv) < 2:
+    print("Usage: python client.py <song1> ...")
+    sys.exit(1)
+
 input_playlist = sys.argv[1:]
 
 data = json.dumps({'songs': input_playlist})
diff --git a/ler.py b/ler.py
new file mode 100644
index 0000000..320c011
--- /dev/null
+++ b/ler.py
@@ -0,0 +1,13 @@
+import pandas as pd
+from mlxtend.frequent_patterns import apriori
+from mlxtend.frequent_patterns import association_rules
+
+df_spotify_ds1 = pd.read_csv("/home/datasets/2023_spotify_ds1.csv")
+df_spotify_ds2 = pd.read_csv("/home/datasets/2023_spotify_ds2.csv")
+df_songs = pd.read_csv("/home/datasets/2023_spotify_songs.csv")
+
+
+
+print(f"spotify_ds1:\n{df_spotify_ds1.head()}")
+print(f"spotify_ds2:\n{df_spotify_ds2.head()}")
+print(f"songs:\n{df_songs.head()}") 
\ No newline at end of file