v2.0.1

• added features description to readme
• added shortcuts to read me
• Fixed missing wakeword models download

v2.0.0

Wakeword detection + Local TTS

• Added wake word detection (call "sprite" and then say your request)
• Added "listening" alert sound before starting microphone audio transcription
• Added local TTS
• Removed the need for writing a prompt template for each model
• Added automatic download of huggingface model when defined on settings file (llamacpp backend only)
• Ollama backed now defaults to local host if no url is passed
• Replaced default model with a llama3 based one
• Fixed memory loading

Support for STT using CPU

• Trancription will now work on cpu