Lokutor

Text-to-speech service implementation using Lokutor’s CPU-only speech synthesis

Overview

LokutorTTSService generates speech from text using Lokutor’s text-to-speech API over a persistent WebSocket connection. Lokutor runs on CPU-only infrastructure and supports multiple languages (EN, ES, FR, PT, KO). The service outputs TTSAudioRawFrames.

Source Repository

Source code, examples, and issues for the Lokutor integration

PyPI Package

The pipecat-lokutor package on PyPI

Lokutor App

Get an API key and test available voices

Documentation

Lokutor API and service documentation

Installation

This is a community-maintained package distributed separately from pipecat-ai:

uv add pipecat-lokutor

Prerequisites

Lokutor Account Setup

Before using the Lokutor TTS service, you need:

1. Lokutor Account: Sign up and get an API key at app.lokutor.com
2. API Key: Used to authenticate the WebSocket connection

Required Environment Variables

• LOKUTOR_API_KEY: Your Lokutor API key (used by the example to populate api_key)

Configuration

api_key · str · required

Lokutor API key used to authenticate the WebSocket connection.

voice_id · str · default: "F1"

Identifier of the voice to use. Browse and test voices at app.lokutor.com.

sample_rate · int · default: "44100"

Output audio sample rate in Hz.

params · LokutorTTSService.InputParams · default: "None"

Runtime input parameters for generation. See Input Parameters below.

settings · LokutorTTSSettings · default: "None"

Optional Pipecat TTS settings applied on top of the defaults.

base_url · str · default: "wss://api.lokutor.com/ws"

Base WebSocket URL for the Lokutor API.

Input Parameters

Runtime-configurable parameters passed via the params constructor argument using LokutorTTSService.InputParams(...).

Parameter	Type	Default	Description
language	Language	None	Synthesis language (EN, ES, FR, PT, KO).
speed	float	1.0	Speech speed multiplier.
steps	int	5	Number of synthesis steps.
visemes	bool	False	Whether to request viseme data.

Available parameters and defaults are defined by the integration. See the source repository for the authoritative, up-to-date list.

Usage

from pipecat_lokutor import LokutorTTSService

tts = LokutorTTSService(
    api_key="your_api_key",
    voice_id="F1",
    params=LokutorTTSService.InputParams(
        language="en",
        speed=1.0,
        steps=5,
        visemes=False,
    ),
)

# ... add tts to your pipeline:
from pipecat.pipeline.pipeline import Pipeline

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

Compatibility

Tested with Pipecat v0.0.86+. Check the source repository for the latest tested version and changelog.