Gives Auto-GPT a voice!

The AI now spesks it's thoughts every turn.

AutonomousAI/data/prompt.txt

@@ -56,6 +56,7 @@ RESPONSE FORMAT:
 "reasoning": "reasoning",
 "plan": "short bulleted long-term plan",
 "criticism": "constructive self-criticism"
+"speak": "thoughts summary to say to user"
 }
 }
 

AutonomousAI/main.py

@@ -7,6 +7,7 @@ import chat
 from colorama import Fore, Style
 from spinner import Spinner
 import time
+import speak
 
 def print_to_console(title, title_color, content, min_typing_speed=0.05, max_typing_speed=0.01):
     print(title_color + title + " " + Style.RESET_ALL, end="")
@@ -35,11 +36,13 @@ def print_assistant_thoughts(assistant_reply):
             assistant_thoughts_reasoning = assistant_thoughts.get("reasoning")
             assistant_thoughts_plan = assistant_thoughts.get("plan")
             assistant_thoughts_criticism = assistant_thoughts.get("criticism")
+            assistant_thoughts_speak = assistant_thoughts.get("speak")
         else:
             assistant_thoughts_text = None
             assistant_thoughts_reasoning = None
             assistant_thoughts_plan = None
             assistant_thoughts_criticism = None
+            assistant_thoughts_speak = None
         
         print_to_console(f"{ai_name.upper()} THOUGHTS:", Fore.YELLOW, assistant_thoughts_text)
         print_to_console("REASONING:", Fore.YELLOW, assistant_thoughts_reasoning)
@@ -57,6 +60,10 @@ def print_assistant_thoughts(assistant_reply):
                     print_to_console("- ", Fore.GREEN, line.strip())
         print_to_console("CRITICISM:", Fore.YELLOW, assistant_thoughts_criticism)
 
+        # Speak the assistant's thoughts
+        if assistant_thoughts_speak:
+            speak.say_text(assistant_thoughts_speak)
+
     except json.decoder.JSONDecodeError:
         print_to_console("Error: Invalid JSON\n", Fore.RED, assistant_reply)
     # All other errors, return "Error: + error message"

AutonomousAI/speak.py

@@ -0,0 +1,29 @@
+import os
+from playsound import playsound
+import requests
+import keys
+
+voice_id = "ErXwobaYiN019PkySvjV"
+tts_url = "https://api.elevenlabs.io/v1/text-to-speech/{voice_id}".format(voice_id=voice_id)
+
+tts_headers = {
+    "Content-Type": "application/json",
+    "xi-api-key": keys.ELEVENLABS_API_KEY
+}
+
+
+def say_text(text):
+    formatted_message = {"text": text}
+    response = requests.post(
+        tts_url, headers=tts_headers, json=formatted_message)
+
+    if response.status_code == 200:
+        with open("speech.mpeg", "wb") as f:
+            f.write(response.content)
+        playsound("speech.mpeg")
+        # Delete audio file
+        os.remove("speech.mpeg")
+    else:
+        print("Request failed with status code:", response.status_code)
+        print("Response content:", response.content)
+