Cartesia

software/main.py

@@ -258,7 +258,7 @@ def run(
 
         ### START LIVEKIT WORKER
         if server == "livekit":
-            time.sleep(5)
+            time.sleep(1)
             # These are needed to communicate with the worker's entrypoint
             os.environ['INTERPRETER_SERVER_HOST'] = light_server_host
             os.environ['INTERPRETER_SERVER_PORT'] = str(light_server_port)
@@ -274,6 +274,7 @@ def run(
             )).to_jwt())
 
             # meet_url = f'http://localhost:3000/custom?liveKitUrl={url.replace("http", "ws")}&token={token}\n\n'
+            meet_url = f'https://meet.livekit.io/custom?liveKitUrl={url.replace("http", "ws")}&token={token}\n\n'
             print("\n")
             print("For debugging, you can join a video call with your assistant. Click the link below, then send a chat message that says {CONTEXT_MODE_OFF}, then begin speaking:")
             print(meet_url)

software/pyproject.toml

@@ -26,6 +26,7 @@ realtimestt = "^0.2.41"
 pynput = "^1.7.7"
 yaspin = "^3.0.2"
 pywebview = "^5.2"
+livekit-plugins-cartesia = "^0.4.2"
 
 [build-system]
 requires = ["poetry-core"]

software/source/server/livekit/worker.py

@@ -5,7 +5,7 @@ from livekit.agents import AutoSubscribe, JobContext, WorkerOptions, cli
 from livekit.agents.llm import ChatContext, ChatMessage
 from livekit import rtc
 from livekit.agents.voice_assistant import VoiceAssistant
-from livekit.plugins import deepgram, openai, silero, elevenlabs
+from livekit.plugins import deepgram, openai, silero, elevenlabs, cartesia
 from dotenv import load_dotenv
 import sys
 import numpy as np
@@ -81,7 +81,7 @@ async def entrypoint(ctx: JobContext):
     elif tts_provider == 'elevenlabs':
         tts = elevenlabs.TTS()
     elif tts_provider == 'cartesia':
-        pass # import plugin, TODO support this
+        tts = cartesia.TTS()
     else:
         raise ValueError(f"Unsupported TTS provider: {tts_provider}. Please set 01_TTS environment variable to 'openai' or 'elevenlabs'.")
 

software/source/server/profiles/default.py

@@ -1,7 +1,7 @@
 from interpreter import AsyncInterpreter
 interpreter = AsyncInterpreter()
 
-interpreter.tts = "elevenlabs" # This should be cartesia once we support it
+interpreter.tts = "cartesia"
 interpreter.stt = "deepgram" # This is only used for the livekit server. The light server runs faster-whisper locally
 
 # Connect your 01 to a language model

software/source/server/profiles/fast.py

@@ -1,7 +1,7 @@
 from interpreter import AsyncInterpreter
 interpreter = AsyncInterpreter()
 
-interpreter.tts = "elevenlabs" # This should be cartesia once we support it
+interpreter.tts = "cartesia" # This should be cartesia once we support it
 interpreter.stt = "deepgram" # This is only used for the livekit server. The light server runs faster-whisper locally
 
 interpreter.llm.model = "gpt-4o-mini"