Skip to main content

​
Bot with ElevenLabs TTS + OpenAI + Deepgram

A full Daily.co bot using ElevenLabs for speech synthesis, OpenAI for the LLM, and Deepgram for STT. 
import asyncio
import os
from dotenv import load_dotenv
from loguru import logger

from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.task import PipelineParams, PipelineTask
from pipecat.processors.aggregators.llm_context import LLMContext
from pipecat.processors.aggregators.llm_response_universal import (
    LLMContextAggregatorPair,
    LLMUserAggregatorParams,
)
from pipecat.services.deepgram.stt import DeepgramSTTService
from pipecat.services.elevenlabs.tts import ElevenLabsTTSService
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.daily.transport import DailyParams, DailyTransport

from agenthuman.api import NewSessionRequest
from agenthuman.video import AgentHumanVideoService

load_dotenv()


async def run_bot(transport):
    stt = DeepgramSTTService(api_key=os.getenv("DEEPGRAM_API_KEY"))

    tts = ElevenLabsTTSService(
        api_key=os.getenv("ELEVENLABS_API_KEY"),
        voice_id="21m00Tcm4TlvDq8ikWAM",
    )

    llm = OpenAILLMService(
        api_key=os.getenv("OPENAI_API_KEY"),
        settings=OpenAILLMService.Settings(
            system_instruction="You are a friendly AI assistant. Respond naturally and keep your answers conversational.",
        ),
    )

    agentHuman = AgentHumanVideoService(
        api_key=os.getenv("AGENTHUMAN_API_KEY"),
        session_request=NewSessionRequest(
            avatar="avat_01KMDFWB9SW1QX4TVP0RT1RFYQ"
        ),
        transport=transport,
    )

    context = LLMContext()
    user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
        context,
        user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
    )

    pipeline = Pipeline([
        transport.input(),
        stt,
        user_aggregator,
        llm,
        tts,
        agentHuman,
        transport.output(),
        assistant_aggregator,
    ])

    task = PipelineTask(
        pipeline,
        params=PipelineParams(
            enable_metrics=True,
            enable_usage_metrics=True,
        ),
    )

    @transport.event_handler("on_client_connected")
    async def on_client_connected(transport, client):
        logger.info("Client connected")
        context.add_message(
            {"role": "user", "content": "Say hello and briefly introduce yourself."}
        )
        await task.queue_frames([LLMRunFrame()])

    @transport.event_handler("on_client_disconnected")
    async def on_client_disconnected(transport, client):
        logger.info("Client disconnected")
        await task.cancel()

    runner = PipelineRunner()
    await runner.run(task)


async def main():
    transport = DailyTransport(
        room_url=os.getenv("DAILY_ROOM_URL"),
        token=os.getenv("DAILY_TOKEN"),
        bot_name="AI Avatar",
        params=DailyParams(
            audio_in_enabled=True,
            audio_out_enabled=True,
            video_out_enabled=True,
            video_out_is_live=True,
            video_out_width=1200,
            video_out_height=1200,
            video_out_bitrate=2_000_000,
        ),
    )
    await run_bot(transport)


if __name__ == "__main__":
    asyncio.run(main())

​
Bot with Cartesia TTS

Swap ElevenLabs for Cartesia — everything else stays the same. 
from pipecat.services.cartesia.tts import CartesiaTTSService

tts = CartesiaTTSService(
    api_key=os.getenv("CARTESIA_API_KEY"),
    settings=CartesiaTTSService.Settings(
        voice="71a7ad14-091c-4e8e-a314-022ece01c121",  # British Reading Lady
    ),
)
Then use the same pipeline and AgentHumanVideoService setup as above.

​
Portrait Avatar (Mobile Layout)

Use a 3:4 aspect ratio for portrait video, suitable for mobile UIs. 
transport = DailyTransport(
    room_url=os.getenv("DAILY_ROOM_URL"),
    token=os.getenv("DAILY_TOKEN"),
    bot_name="AI Avatar",
    params=DailyParams(
        audio_in_enabled=True,
        audio_out_enabled=True,
        video_out_enabled=True,
        video_out_is_live=True,
        video_out_width=900,
        video_out_height=1200,   # portrait dimensions
    ),
)

agentHuman = AgentHumanVideoService(
    api_key=os.getenv("AGENTHUMAN_API_KEY"),
    session_request=NewSessionRequest(
        avatar="avat_01KMDFWB9SW1QX4TVP0RT1RFYQ",
        aspect_ratio="3:4",   # explicit portrait ratio
    ),
    transport=transport,
)

​
Using Pipecat Cloud / create_transport

When deploying to Pipecat Cloud or using the pipecat.runner helpers, use create_transport to handle transport creation from RunnerArguments: 
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.base_transport import TransportParams

async def bot(runner_args: RunnerArguments):
    transport_params = {
        "daily": lambda: DailyParams(
            audio_in_enabled=True,
            audio_out_enabled=True,
            video_out_enabled=True,
            video_out_is_live=True,
            video_out_width=1200,
            video_out_height=1200,
            video_out_bitrate=2_000_000,
        ),
        "webrtc": lambda: TransportParams(
            audio_in_enabled=True,
            audio_out_enabled=True,
            video_out_enabled=True,
            video_out_is_live=True,
            video_out_width=1200,
            video_out_height=1200,
        ),
    }

    transport = await create_transport(runner_args, transport_params)
    await run_bot(transport, runner_args)


if __name__ == "__main__":
    from pipecat.runner.run import main
    main()