Korbinian Maag | Artibex

🎙️ Piper HTTP 🎙️

Dockerized HTTP service for Piper Text-to-Speech

Overview 📖

A Docker image that runs the Piper HTTP service from the Rhasspy project. This containerized solution provides a simple HTTP API for high-quality neural text-to-speech synthesis, with automatic voice model downloads and easy deployment. No complex setup required - just pull the Docker image and start converting text to speech.

Features 🚀

• Fully containerized Piper TTS HTTP service
• Automatic voice model downloads from Hugging Face
• Simple REST API for text-to-speech conversion
• Multi-language voice support
• Available on Docker Hub for instant deployment
• Includes Python client example for easy integration

Quick Start - Server Setup 🚀

Build from source:

Build from source

$ docker build -t piper .
$ docker run --name piper -p 5000:5000 piper

Or use pre-built image from Docker Hub:

Docker Hub (pre-built image)

$ docker run -e MODEL_DOWNLOAD_LINK=https://huggingface.co/rhasspy/piper-voices/resolve/v1.0.0/en/en_US/kusal/medium/en_US-kusal-medium.onnx?download=true --name piper -p 5000:5000 artibex/piper-http

Quick Start - Client Usage 🎯

Use the included Python client script to send text to the Piper service and play the generated audio:

Install dependencies:

Install Python dependencies

$ pip install requests playsound

Python client example (piper-tts.py):

piper-tts.py

import os
import sys
import requests
import random
from playsound import playsound

# Take textToSpeak as argument
if len(sys.argv) < 2:
    print("Usage: python piper-tts.py \"Hello World!\"")
    sys.exit(1)

textToSpeak = sys.argv[1]
urlPiper = "http://localhost:5000"

# Create random filename for output
outputFilename = "output" + str(random.randint(1,1000000)) + ".wav"

# Create the payload
payload = {'text': textToSpeak}

# Send request to piper
try:
    r = requests.get(urlPiper, params=payload)
    r.raise_for_status()
except requests.exceptions.RequestException as e:
    print(f"Could not send request to piper: {e}")
    sys.exit(1)

# Save the response to a file
with open(outputFilename, 'wb') as fd:
    for chunk in r.iter_content(chunk_size=128):
        fd.write(chunk)

# Play the audio
playsound(outputFilename)

# Clean up
os.remove(outputFilename)

Usage examples:

Usage examples

$ python piper-tts.py "Hello World!"
$ python piper-tts.py "Welcome to Piper TTS"