Swarmauri PlayHT Text-to-Speech

swarmauri_tts_playht is the standalone PlayHT provider package for Swarmauri. Its PlayHTModel implements TTSBase and converts independent text inputs into audio files without pretending to be an LLM.

Features

• Native TTSBase provider registered as swarmauri.tts.PlayHTModel.
• Synchronous, asynchronous, and batch audio generation.
• PlayHT voice discovery and cloned-voice management.
• Configurable voice engine, voice, output format, and timeout.
• Agent-only CLI support for positional, piped, file-based, and interactive text.
• Command-line selection of the PlayHT voice model and voice.

Installation

uv add swarmauri_tts_playht

pip install swarmauri_tts_playht

Set PLAYHT_API_KEY and PLAYHT_USER_ID before generating audio.

Usage

import os

from swarmauri_tts_playht import PlayHTModel

tts = PlayHTModel(
    api_key=os.environ["PLAYHT_API_KEY"],
    user_id=os.environ["PLAYHT_USER_ID"],
    name="Play3.0-mini",
    voice="Adolfo",
)
audio_path = tts.predict("Welcome to Swarmauri.", "welcome.mp3")
print(audio_path)

Generate through the stateless TextToSpeechAgent:

swarmauri-playht "Welcome to Swarmauri." \
  --voice-model PlayHT2.0-turbo \
  --voice Adolfo \
  --output welcome.mp3

Every CLI invocation uses TextToSpeechAgent. It adds no conversation, memory, system context, or text-generating LLM. --model remains available as an alias for --voice-model.

The supported command-line voice engines are Play3.0-mini, PlayHT2.0-turbo, PlayHT1.0, and PlayHT2.0. Text can also come from a UTF-8 file or standard input:

swarmauri-playht --prompt-file announcement.txt --output announcement.mp3
echo "Welcome to Swarmauri." | swarmauri-playht --output welcome.mp3

Use the same stateless workflow from Python:

import os

from swarmauri_agent_texttospeech import TextToSpeechAgent
from swarmauri_tts_playht import PlayHTModel

tts = PlayHTModel(
    api_key=os.environ["PLAYHT_API_KEY"],
    user_id=os.environ["PLAYHT_USER_ID"],
    name="PlayHT2.0-turbo",
    voice="Adolfo",
)
agent = TextToSpeechAgent(tts=tts, output_path="welcome.mp3")
audio_path = agent.exec("Welcome to Swarmauri.")

Contributing

This package is part of the Swarmauri SDK. Contributions should follow the repository contribution and style guides.

License

Apache-2.0