v1.6.2 Update
This update introduces new optional features, performance improvements, and additional user controls.
New Features
Optional Cloud LLM Support
Added support for external cloud-based language models, including OpenRouter, OpenAI, and Anthropic providers. This feature is completely optional and requires a user-provided API key. Add your key and press “Fetch Models” to select which model you would like to use. The application remains fully functional offline.
Optional Cloud Voice (ElevenLabs)
Added support for external voice synthesis via ElevenLabs. This feature is optional, requires a user-provided API key, and is disabled by default. When creating an API key, please ensure that text-to-speech permissions are enabled and that voice access permissions are granted.
Character Unload / Disable Option
You can now disable and unload the character from memory when needed. This is especially useful for lower-end systems where resources are limited, allowing more performance to be allocated to the LLM while reducing distractions.
Important Notes
The application continues to work fully offline with all core features available without any external services. Cloud features are entirely optional and intended for advanced users. Enabling cloud features may send text or audio data to the selected third-party provider, depending on configuration.
More to come soon.
