Kokoro TTS plugin guide | LiveKit Documentation

Available inPython

Overview

Kokoro is an open-weight TTS model you can run locally. Kokoro-FastAPI wraps it in an OpenAI-compatible endpoint, which the OpenAI plugin connects to directly.

Installation

Install the OpenAI plugin to add Kokoro support:

uv add "livekit-agents[openai]~=1.5"

Usage

Run Kokoro-FastAPI, then point openai.TTS at its OpenAI-compatible endpoint. For example, you can use this TTS in the Voice AI quickstart:

from livekit.plugins import openai

session = AgentSession(
    tts=openai.TTS(
        model="kokoro",
        voice="af_alloy",
        api_key="not-needed",
        base_url="http://localhost:8880/v1",
        response_format="wav",
    ),
    # ... llm, stt, vad, turn_detection, etc.
)

Parameters

This section describes the parameters used to connect to a Kokoro-FastAPI server. See the plugin reference for a complete list of all available parameters.

modelstring

Model to use for speech generation. Set to kokoro to use the Kokoro model served by Kokoro-FastAPI.

voicestring

Kokoro voice name, such as af_alloy. For the full list of available voices, see the Kokoro-FastAPI documentation .

base_urlstring

URL of your Kokoro-FastAPI server's OpenAI-compatible endpoint. Kokoro-FastAPI listens on http://localhost:8880/v1 by default.

api_keystring

Kokoro-FastAPI doesn't require authentication, but the OpenAI client rejects an empty value. Set any non-empty placeholder, such as not-needed.

response_formatstring

Audio format returned by the server. Kokoro-FastAPI supports several formats, such as wav.

Additional resources

The following resources provide more information about using Kokoro with LiveKit Agents.

Kokoro-FastAPI

Run Kokoro locally behind an OpenAI-compatible endpoint.

OpenAI TTS plugin

The OpenAI TTS plugin that powers this integration.

Python plugin reference

Reference for the OpenAI TTS plugin.

Voice AI quickstart

Get started with LiveKit Agents and Kokoro.