Skip to content

v2.0.0 - ElevenLabs TTS Integration

Latest
Latest

Choose a tag to compare

View all tags
@binaryninja binaryninja released this 19 Jul 14:35
· 3 commits to main since this release
v2.0.0
70384d4

🎤 Major TTS Integration Update

What's New

ElevenLabs as Default TTS Provider

HoldTranscribe now uses ElevenLabs as the default text-to-speech engine, providing professional-quality voice synthesis with minimal setup requirements.

✨ Key Features

  • 🗣️ High-Quality Voices: Professional-grade synthesis via ElevenLabs API
  • ⚡ Fast Synthesis: Multiple models for different speed/quality needs
  • 🌊 Streaming Support: Real-time audio streaming for low latency
  • ☁️ Cloud-Based: No local GPU required, just an API key
  • 🎭 Voice Customization: Configurable voice parameters
  • 🔧 Easy Setup: 5-minute setup with API key

🚀 Quick Start

  1. Get API key: https://elevenlabs.io/app/settings/api-keys
  2. Install: pip install elevenlabs>=1.0.0
  3. Set key: export ELEVENLABS_API_KEY="your_key"
  4. Run: python -m holdtranscribe.main --tts

📊 Model Options

  • eleven_multilingual_v2 - High quality (default)
  • eleven_turbo_v2_5 - Faster synthesis
  • eleven_flash_v2_5 - Fastest synthesis

🔄 Backward Compatibility

  • Moshi TTS still available via --tts-model moshi
  • All existing APIs maintained
  • Previous configurations continue to work

📚 New Documentation

  • ELEVENLABS_QUICKSTART.md - 5-minute setup guide
  • TTS_SETUP.md - Comprehensive documentation
  • setup_elevenlabs.py - Interactive setup script
  • demo_elevenlabs_integration.py - Feature demonstration

🎯 Usage Examples

# Basic TTS (ElevenLabs default)
holdtranscribe --tts

# AI Assistant with TTS
holdtranscribe --model mistralai/Voxtral-Mini-3B-2507 --tts

# Fast synthesis
holdtranscribe --tts --tts-model eleven_turbo_v2_5

🔧 Testing

# Test integration
python test_elevenlabs_integration.py

# Run comprehensive demo
python demo_elevenlabs_integration.py

# Interactive setup
python setup_elevenlabs.py

⚠️ Important Fix (v2.0.0)

Fixed missing model files: The initial release missed core model implementation files due to an overly broad .gitignore rule. This has been corrected and all model source files are now properly included:

  • whisper_model.py - Whisper transcription implementation
  • voxtral_model.py - Voxtral AI assistant implementation
  • moshi_model.py - Moshi TTS implementation
  • dia_model.py - DIA TTS implementation
  • elevenlabs_model.py - ElevenLabs TTS core implementation
  • elevenlabs_wrapper.py - ElevenLabs TTS wrapper for HoldTranscribe

This release significantly improves voice synthesis quality while maintaining the flexibility to use alternative TTS providers.

Full Changelog: a7088e8...494b16b

Assets 2
Loading