Skip to main content
Build a voice agent using Deepgram for speech recognition, OpenAI GPT-4o for conversation, and Rime for text-to-speech synthesis. Best for: Applications requiring precise pronunciation control, custom phonemes, and word-level timing.

​
Prerequisites

ServiceWhat You Need
PlivoAuth ID, Auth Token, Voice-enabled phone number
DeepgramAPI key from console.deepgram.com
OpenAIAPI key from platform.openai.com
RimeAPI key from rime.ai

​
Installation

pip install "pipecat-ai[deepgram,openai,rime]"

​
Environment Variables

# Plivo credentials
PLIVO_AUTH_ID=your_auth_id
PLIVO_AUTH_TOKEN=your_auth_token
PLIVO_PHONE_NUMBER=+1234567890

# AI service credentials
DEEPGRAM_API_KEY=your_deepgram_key
OPENAI_API_KEY=sk-your_openai_key
RIME_API_KEY=your_rime_key

​
Pipeline Configuration

from pipecat.services.deepgram import DeepgramSTTService
from pipecat.services.openai import OpenAILLMService
from pipecat.services.rime import RimeTTSService

# Speech-to-Text
stt = DeepgramSTTService(
    api_key=os.getenv("DEEPGRAM_API_KEY"),
)

# Language Model
llm = OpenAILLMService(
    api_key=os.getenv("OPENAI_API_KEY"),
    model="gpt-4o",
)

# Text-to-Speech
tts = RimeTTSService(
    api_key=os.getenv("RIME_API_KEY"),
    # voice="your_voice_id",
)

​
Service Details

​
Deepgram STT

Real-time speech recognition with interim results and language detection.
OptionDescription
DeepgramSTTServiceStandard WebSocket transcription
DeepgramFluxSTTServiceEnhanced turn detection for conversations

​
OpenAI LLM

Chat completion with GPT-4o supporting streaming responses and function calling.
ModelDescription
gpt-4oMost capable, multimodal
gpt-4o-miniFaster, cost-effective

​
Rime TTS

Real-time voice synthesis with word-level timing and precise pronunciation control.
FeatureMethod
Spell out textSPELL("ABC")
Insert pausePAUSE_TAG(0.5)
Custom pronunciationPRONOUNCE(text, word, phoneme)
Adjust speed inlineINLINE_SPEED(text, 1.2)
Service options:
  • RimeTTSService - WebSocket-based, real-time with word timestamps
  • RimeHttpTTSService - HTTP-based, simpler setup

​
Pronunciation Control

Rime excels at precise pronunciation control for names, technical terms, and branded content.

​
Custom Pronunciations

from pipecat.services.rime import PRONOUNCE

# Replace word with phoneme pronunciation
text = PRONOUNCE(
    "Welcome to Plivo",
    "Plivo",
    "plee-voh"
)

​
Spelling Out Text

from pipecat.services.rime import SPELL

# Spell out acronyms or codes
text = SPELL("API")  # Says "A P I"

​
Dynamic Speed Control

from pipecat.services.rime import INLINE_SPEED

# Speed up specific sections
text = INLINE_SPEED("Terms and conditions apply", 1.3)

​
Quick Start

​
Inbound Calls

git clone https://github.com/pipecat-ai/pipecat-examples.git
cd pipecat-examples/plivo-chatbot/inbound

# Configure environment
cp env.example .env
# Edit .env with your credentials

# Start server
uv sync && uv run server.py

# Expose with ngrok (development)
ngrok http 7860
Configure your Plivo number’s Answer URL to your ngrok URL.

​
Outbound Calls

cd pipecat-examples/plivo-chatbot/outbound

cp env.example .env
uv sync && uv run server.py

# Initiate a call
curl -X POST http://localhost:7860/start \
  -H "Content-Type: application/json" \
  -d '{"phone_number": "+1234567890"}'

​
When to Use Rime

Choose Rime when:
  • You need precise control over pronunciation
  • Your content includes technical terms, names, or branded words
  • You want word-level timing for synchronized experiences
  • You need inline speed adjustments
Choose Cartesia or ElevenLabs when:
  • You need emotion/expression controls
  • You want voice cloning capabilities
  • You need broader multilingual support