persian-speech 1.0.1

Natural Persian Text-to-Speech. One command.

Persian Speech

Natural Persian Text-to-Speech. One command.

Install

pip install persian-speech
python -c "import urllib.request, os; os.makedirs('models', exist_ok=True); urllib.request.urlretrieve('https://huggingface.co/MahtaFetrat/Mana-Persian-Piper/resolve/main/fa_IR-mana-medium.onnx', 'models/fa_IR-mana-medium.onnx'); urllib.request.urlretrieve('https://huggingface.co/MahtaFetrat/Mana-Persian-Piper/resolve/main/fa_IR-mana-medium.onnx.json', 'models/fa_IR-mana-medium.onnx.json'); print('Done')"

Usage

Command line:

speech "سلام به همگی" output.wav

Python:

from persian_speech_runtime import talk
talk("سلام به همگی", "output.wav")

Server:

python -c "from persian_speech_runtime import PiperSpeechEngine; from http.server import HTTPServer, BaseHTTPRequestHandler; from urllib.parse import urlparse, parse_qs; tts = PiperSpeechEngine(); exec('class H(BaseHTTPRequestHandler):\\n def do_GET(self):\\n  q = parse_qs(urlparse(self.path).query)\\n  t = q.get(\"text\",[\"سلام\"])[0]\\n  self.send_response(200)\\n  self.send_header(\"Content-Type\",\"audio/wav\")\\n  self.end_headers()\\n  self.wfile.write(tts.stream(t))'); HTTPServer(('0.0.0.0',8000), H).serve_forever()"
# http://localhost:8000/tts?text=سلام

Model

Built on Mana-Persian-Piper, trained on 114 hours of Persian speech from the Mana-TTS dataset.

API

from persian_speech_runtime import PiperSpeechEngine, talk

# One-liner
talk("سلام به همگی", "output.wav")

# Engine
tts = PiperSpeechEngine()
tts.save("سلام به همگی", "output.wav")   # Save to file
tts.stream("سلام به همگی")               # Get WAV bytes
tts.synthesize("سلام به همگی")           # Get numpy array

License

MIT