SelenaCore Update: NVIDIA Jetson Support and Advanced LLM Orchestration

I am excited to announce a major leap in SelenaCore's local processing capabilities. Our latest update focuses on maximizing hardware utilization, specifically for edge computing environments. We have implemented automatic CUDA and GPU detection 667e08c, which now intelligently configures engines based on your hardware profile.

For users on NVIDIA Jetson platforms, we've introduced dedicated GPU containers for both Piper TTS 9cc251a and Ollama 2e3e784, ensuring low-latency performance. To keep you informed, the UI now features acceleration badges (CUDA/CPU/Cloud) for each engine 6ef3ef5.

On the LLM front, we have added a llama.cpp provider and a model catalog equipped with a RAM filter to prevent system overloads a0c9568. Additionally, developers can now fine-tune AI behavior via a new UI-based system prompt editor bbeaa0a. To improve synthesis quality, we've also standardized text processing by normalizing AI responses to lowercase before they reach the TTS engine 3d97309.

Detailed technical guides are now available in our Voice Settings Documentation.