voice-of-python 0.1.0

Offline multilingual voice bot with Vosk ASR and pyttsx3 TTS

voice_of_python

Offline Multilingual Voice Bot Library for Python

voice_of_python is a Python library that provides offline multilingual speech recognition and text-to-speech (TTS) capabilities. It uses the Vosk toolkit for offline speech-to-text and pyttsx3 for offline TTS, supporting multiple languages through automatic language detection.

---

Features

• Record audio from microphone or accept existing WAV audio files.
• Offline speech-to-text using Vosk models (no API or internet required).
• Automatic language detection with langid.
• Text-to-speech in detected language voice (if available locally).
• Works fully on CPU; suitable for local and privacy-sensitive applications.
• Designed for easy integration with chatbots and LLMs.

---

Installation

Install dependencies:

pip install vosk langid pyttsx3 sounddevice numpy soundfile

Note: Download and provide local Vosk models for supported languages separately.

---

Usage

1. Initialize the library with Vosk model paths

from voice_of_python import MultiLingualVoiceBot

model_paths = {
"en": "/path/to/vosk-model-en-us-0.22",
"hi": "/path/to/vosk-model-small-hi-0.22"
}

bot = MultiLingualVoiceBot(model_paths)

2. Record from microphone and get transcription

text, lang_code = bot.record_and_transcribe(record_duration=5)
print(f"Detected language: {lang_code}")
print(f"Recognized text: {text}")

3. Integrate with your LLM or AI model

Pass text to your backend model to generate a reply message.

reply_text = "यहाँ आपका उत्तर है"

4. Speak reply in detected language

bot.speak_text(reply_text, lang_code)

5. Alternatively, process external audio file and speak reply

audio_file = "user.wav" reply_text = "Thank you for your question." bot.process_audio_and_reply(audio_file, reply_text)

---

API Reference

• MultiLingualVoiceBot(model_paths: dict, sample_rate: int = 16000)
  Initialize with {language_code: vosk_model_path} mapping.

• record_and_transcribe(record_duration: int = 5) -> (str, str)
  Record from mic and return transcription and detected language code.

• process_audio_and_reply(audio_input: Union[str, np.ndarray], reply_text: str)
  Transcribe audio (file path or raw numpy array), detect language, speak reply text.

• speak_text(text: str, lang_code: str)
  Speak given text in specified language voice.

---

Contributing

Contributions welcome! Please open issues or pull requests on GitHub.

---

Acknowledgements

• Vosk: https://alphacephei.com/vosk/
• pyttsx3: https://pyttsx3.readthedocs.io/en/latest/
• langid: https://github.com/saffsd/langid.py

---

Support

Please report issues or questions on the GitHub repository.

---

Enjoy building multilingual, offline voice-enabled applications with voice_of_python!