🎙️ Auretrix

Unified Speech Recognition Library - Simplify speech-to-text with multi-provider support, auto-fallback, and Indian language focus.

---

🚀 Features

• 🔄 Multi-Provider Support - Whisper, Google Speech, AssemblyAI, Azure
• 🛡️ Auto-Fallback - Automatically switches providers on failure
• 💰 Cost Optimization - Intelligently chooses cheapest provider
• 🇮🇳 Indian Languages - First-class support for Hindi, Tamil, Telugu, and 10+ Indian languages
• 📝 Multiple Formats - Export to TXT, JSON, SRT, VTT
• ⚡ Simple API - One-liner transcription
• 🔧 Flexible - Use local models or cloud APIs

---

📦 Installation

# Basic installation
pip install auretrix

# With Whisper support (local transcription)
pip install auretrix[whisper]

# With all providers
pip install auretrix[all]

---

🎯 Quick Start

Simple Transcription

from auretrix import transcribe_audio

# One-liner transcription
text = transcribe_audio('meeting.mp3', language='hi-IN')
print(text)

Advanced Usage

from auretrix import SpeechRecognizer

# Initialize with multiple providers
recognizer = SpeechRecognizer(
    providers=['whisper', 'google', 'assemblyai'],
    auto_fallback=True,
    optimize_cost=True
)

# Transcribe with detailed results
result = recognizer.recognize('podcast.mp3', language='en-US')

print(f"Text: {result['text']}")
print(f"Provider: {result['provider']}")
print(f"Confidence: {result['confidence']}")

Batch Transcription

from auretrix import batch_transcribe

files = ['audio1.mp3', 'audio2.wav', 'audio3.m4a']
results = batch_transcribe(files, language='hi-IN')

for result in results:
    print(f"{result['file']}: {result['text']}")

Export to Subtitles

from auretrix import transcribe_file

# Generate SRT subtitles
result = transcribe_file(
    'video.mp4',
    output_format='srt',
    language='en-US',
    include_timestamps=True
)

print(f"Subtitles saved to: {result['output_file']}")

---

🌏 Language Support

Indian Languages

from auretrix.languages import INDIAN_LANGUAGES, list_indian_languages

# See all supported Indian languages
languages = list_indian_languages()
for lang in languages:
    print(f"{lang['name']} ({lang['native']}): {lang['code']}")

Supported Indian Languages:

• 🇮🇳 Hindi (हिन्दी) - hi-IN
• 🇮🇳 Bengali (বাংলা) - bn-IN
• 🇮🇳 Telugu (తెలుగు) - te-IN
• 🇮🇳 Marathi (मराठी) - mr-IN
• 🇮🇳 Tamil (தமிழ்) - ta-IN
• 🇮🇳 Gujarati (ગુજરાતી) - gu-IN
• 🇮🇳 Kannada (ಕನ್ನಡ) - kn-IN
• 🇮🇳 Malayalam (മലയാളം) - ml-IN
• 🇮🇳 Punjabi (ਪੰਜਾਬੀ) - pa-IN
• 🇮🇳 Urdu (اردو) - ur-IN
• And more...

---

🔑 Provider Setup

Whisper (Local - Free)

from auretrix import SpeechRecognizer

# Use local Whisper model (no API key needed)
recognizer = SpeechRecognizer(
    providers=['whisper'],
    use_local=True,
    model='base'  # Options: tiny, base, small, medium, large
)

Google Speech API

export GOOGLE_SPEECH_API_KEY='your-api-key'

recognizer = SpeechRecognizer(providers=['google'])

AssemblyAI

export ASSEMBLYAI_API_KEY='your-api-key'

Azure Speech

export AZURE_SPEECH_KEY='your-key'
export AZURE_SPEECH_REGION='your-region'

---

💡 Use Cases

1. Podcast Transcription

result = transcribe_file(
    'podcast_episode.mp3',
    output_format='txt',
    language='en-US'
)

2. Meeting Notes

recognizer = SpeechRecognizer(providers=['whisper'])
result = recognizer.recognize('team_meeting.wav', language='hi-IN')

3. Video Subtitles

transcribe_file(
    'video.mp4',
    output_format='srt',
    include_timestamps=True
)

4. Multi-Language Content

# Detect and transcribe multiple languages
for audio_file, lang in [('hindi.mp3', 'hi-IN'), ('tamil.mp3', 'ta-IN')]:
    text = transcribe_audio(audio_file, language=lang)
    print(f"{lang}: {text}")

---

🛠️ Configuration

Environment Variables

# Optional API keys
export GOOGLE_SPEECH_API_KEY='...'
export ASSEMBLYAI_API_KEY='...'
export AZURE_SPEECH_KEY='...'
export AZURE_SPEECH_REGION='...'

Provider Priority

recognizer = SpeechRecognizer(
    providers=['whisper', 'google', 'assemblyai'],  # Priority order
    auto_fallback=True  # Auto switch on failure
)

Cost Optimization

recognizer = SpeechRecognizer(
    optimize_cost=True  # Chooses cheapest provider based on audio length
)

---

📊 Provider Comparison

Provider	Cost	Speed	Languages	Local
Whisper	Free	Fast	99+	✅
Google	$0.024/min	Fast	125+	❌
AssemblyAI	$0.00025/sec	Medium	99+	❌
Azure	$1/hour	Fast	100+	❌

---

🤝 Contributing

Contributions are welcome! Check out our Contributing Guide.

---

📄 License

MIT License - see LICENSE file.

---

🔗 Links

• Website: auretrix.com
• Documentation: docs.auretrix.com
• GitHub: github.com/auretrix/auretrix
• PyPI: pypi.org/project/auretrix

---

💬 Support

• 📧 Email: contact@auretrix.com
• 💬 Discord: Join our community
• 🐛 Issues: GitHub Issues

---

Made with ❤️ by Auretrix AI

Building AI tools for creators, by creators.