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


Installation


Environment Variables


Pipeline Configuration


Service Details

Deepgram STT

Real-time speech recognition with interim results and language detection.

OpenAI LLM

Chat completion with GPT-4o supporting streaming responses and function calling.

Rime TTS

Real-time voice synthesis with word-level timing and precise pronunciation control. 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

Spelling Out Text

Dynamic Speed Control


Quick Start

Inbound Calls

Configure your Plivo number’s Answer URL to your ngrok URL.

Outbound Calls


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