Merge pull request #20 from livekit-examples/use-turn-detector

enable turn-detector by default

Author: davidzhao

.env.example

@@ -3,3 +3,4 @@ LIVEKIT_API_KEY=<your API Key>
 LIVEKIT_API_SECRET=<your API Secret>
 OPENAI_API_KEY=<To use other providers, press Enter for now and edit .env.local>
 DEEPGRAM_API_KEY=<To use other providers, press Enter for now and edit .env.local>
+CARTESIA_API_KEY=<To use other providers, press Enter for now and edit .env.local>

README.md

@@ -25,6 +25,7 @@ cd voice-pipeline-agent-python
 python3 -m venv venv
 source venv/bin/activate
 pip install -r requirements.txt
+python3 agent.py download-files
 ```
 
 Set up the environment by copying `.env.example` to `.env.local` and filling in the required values:
@@ -33,6 +34,7 @@ Set up the environment by copying `.env.example` to `.env.local` and filling in
 - `LIVEKIT_API_KEY`
 - `LIVEKIT_API_SECRET`
 - `OPENAI_API_KEY`
+- `CARTESIA_API_KEY`
 - `DEEPGRAM_API_KEY`
 
 You can also do this automatically using the LiveKit CLI:

agent.py

@@ -8,9 +8,10 @@
     WorkerOptions,
     cli,
     llm,
+    metrics,
 )
 from livekit.agents.pipeline import VoicePipelineAgent
-from livekit.plugins import openai, deepgram, silero
+from livekit.plugins import cartesia, openai, deepgram, silero, turn_detector
 
 
 load_dotenv(dotenv_path=".env.local")
@@ -38,18 +39,30 @@ async def entrypoint(ctx: JobContext):
     participant = await ctx.wait_for_participant()
     logger.info(f"starting voice assistant for participant {participant.identity}")
 
-    # This project is configured to use Deepgram STT, OpenAI LLM and TTS plugins
-    # Other great providers exist like Cartesia and ElevenLabs
+    # This project is configured to use Deepgram STT, OpenAI LLM and Cartesia TTS plugins
+    # Other great providers exist like Cerebras, ElevenLabs, Groq, Play.ht, Rime, and more
     # Learn more and pick the best one for your app:
     # https://docs.livekit.io/agents/plugins
     agent = VoicePipelineAgent(
         vad=ctx.proc.userdata["vad"],
         stt=deepgram.STT(),
         llm=openai.LLM(model="gpt-4o-mini"),
-        tts=openai.TTS(),
+        tts=cartesia.TTS(),
+        turn_detector=turn_detector.EOUModel(),
+        # minimum delay for endpointing, used when turn detector believes the user is done with their turn
+        min_endpointing_delay=0.5,
+        # maximum delay for endpointing, used when turn detector does not believe the user is done with their turn
+        max_endpointing_delay=5.0,
         chat_ctx=initial_ctx,
     )
 
+    usage_collector = metrics.UsageCollector()
+
+    @agent.on("metrics_collected")
+    def on_metrics_collected(agent_metrics: metrics.AgentMetrics):
+        metrics.log_metrics(agent_metrics)
+        usage_collector.collect(agent_metrics)
+
     agent.start(ctx.room, participant)
 
     # The agent should be polite and greet the user when it joins :)

requirements.txt

@@ -1,5 +1,7 @@
-livekit-agents>=0.12.1
-livekit-plugins-openai>=0.10.9
-livekit-plugins-deepgram>=0.6.13
+livekit-agents>=0.12.11
+livekit-plugins-openai>=0.10.17
+livekit-plugins-cartesia>=0.4.7
+livekit-plugins-deepgram>=0.6.17
 livekit-plugins-silero>=0.7.4
+livekit-plugins-turn-detector>=0.4.0
 python-dotenv~=1.0

taskfile.yaml

@@ -14,6 +14,7 @@ tasks:
       - platforms: [windows]
         cmd: echo -e "\tpowershell venv/Scripts/Activate.ps1\r"
       - echo -e "\tpip install -r requirements.txt\r"
+      - echo -e "\tpython3 agent.py download-files\r"
       - echo -e "\tpython3 agent.py dev\r\n"
 
   install: