SelenaCore Update: High-Performance Voice Intelligence and Hardware-Accelerated STT

In our latest update, SelenaCore takes a massive leap forward in making voice interaction more fluid and professional. We have introduced hardware-accelerated Speech-to-Text (STT) specifically optimized for NVIDIA Jetson devices using a dedicated faster-whisper GPU server 893298e. This ensures lightning-fast response times even on edge hardware.

To improve the user experience, we have implemented an Energy Gate 366896a and advanced Voice Activity Detection (VAD) 35f68b8. These features prevent 'hallucinations' (the AI hearing things in a silent room) and save processing power by skipping transcription when no one is speaking. Additionally, the new blocking speak() function 91841c7 ensures the system finishes its sentence before performing a requested action, making interactions feel more natural.

For power users and developers, we’ve added a Live STT Monitor bf9a3d1 directly in the UI settings, allowing you to see exactly what the system hears in real-time. We have also removed all hardcoded language restrictions, now pulling all configurations directly from the Piper TTS engine 5e12aef, which simplifies multilingual setups. Detailed setup instructions can be found in our documentation.