Latest updates, features, and releases from SelenaCore.

Significant upgrades to the voice ecosystem including a native Piper TTS server, automated GPU/CPU detection for LLMs, and critical memory leak fixes in media playback.

We've introduced a dedicated Prompts tab for real-time editing and expanded multilingual support for intent classification across six languages.

Major enhancement to the voice interaction layer, introducing unified intent classification, Cloud LLM support, and a comprehensive debugging console for developers.

We've overhauled our documentation structure and addressed critical race conditions in the TTS engine to ensure smoother voice interactions and easier onboarding.

SelenaCore now features a full benchmark suite designed to track server load statistics and ensure peak performance across all modules and hardware environments.

We've overhauled the communication layer with a new WebSocket bus and implemented a comprehensive i18n system to support global accessibility.

Introduction of a dynamic prompt system with i18n support, redesigned voice settings UI, and optimized LLM provider interactions for Ollama and Gemini.

Significant improvements to voice processing with a new Vosk pipeline, streaming TTS, and real-time audio monitoring APIs.

Launching the smarthome_sdk, expanding Zigbee support, and hardening API security through integrity-aware IPC and OAuth proxies.

This update introduces automated background management for local AI servers, non-blocking provider switching, and real-time progress tracking for model installations.