livekit-plugins-soniox

What it is and what it does

This package is a plugin that bridges Soniox's speech APIs into the LiveKit Agents framework, enabling agent applications to transcribe audio in real time and synthesize speech responses. It wraps Soniox's WebSocket streaming interfaces for both STT (speech-to-text) and TTS (text-to-speech), allowing developers to drop Soniox components into an AgentSession alongside other LiveKit services like LLMs.

The plugin is designed for low-latency, streaming workflows: STT can run as a standalone transcription service or within an agent session, while TTS tokenizes text chunks from an LLM and sends them to Soniox as they arrive, enabling incremental speech synthesis rather than waiting for full responses. It requires only livekit-agents as a runtime dependency and a valid Soniox API key to authenticate.

Use it for:

• Build a conversational voice agent that transcribes user speech and synthesizes natural responses in real time.
• Add live transcription to a LiveKit room session using Soniox's multilingual STT capabilities.
• Stream LLM-generated text to Soniox TTS for low-latency speech output in interactive applications.
• Integrate Soniox voice synthesis with configurable language, voice, and speed parameters into an existing LiveKit agent.
• Process WebSocket-based audio streams for transcription without managing Soniox connection details directly.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Integrates Soniox speech-to-text and text-to-speech APIs into LiveKit Agents via WebSocket streaming, enabling real-time transcription and speech synthesis in agent applications.

Yes. The package is actively maintained, has no known vulnerabilities, installs with minimal friction, and is permissively licensed. Install it if you are building a LiveKit agent and want to use Soniox for speech services; it is the intended integration point for that use case. You will need a Soniox API key and Python 3.10+.

Install

livekit-plugins-soniox on PyPI

pip install livekit-plugins-soniox

uv add livekit-plugins-soniox

poetry add livekit-plugins-soniox

Installing livekit-plugins-soniox

Before you install

Installation is straightforward with a single pure-Python wheel dependency on livekit-agents. The package is actively maintained with a recent release and no known vulnerabilities.

License in practice

Licensed under Apache-2.0 (permissive), allowing use in commercial and proprietary projects with minimal restrictions.

Quickstart

pip install livekit-plugins-soniox

from livekit.plugins import soniox

session = AgentSession(
    stt=soniox.STT(),
    tts=soniox.TTS(language="en", voice="Maya", speed=1.0)
)

Requires a Soniox API key set in the SONIOX_API_KEY environment variable; Python 3.10 or later.

Verify before relying

• Whether Soniox API credentials are required upfront or can be deferred until runtime.
• Specific latency characteristics for real-time speech synthesis with LLM streaming.
• Supported Soniox voice options and language coverage beyond the documentation links provided.

Package facts

Evidence: livekit_plugins_soniox-1.6.10-py3-none-any.whl

Keywords: audio, livekit, realtime, soniox, speech-to-text, stt, text-to-speech, tts, webrtc

Tags

More Artificial Intelligence packages

Further reading