Tags: hurdcog/Layla-Server

feat: add backend probing functionality for OpenAI-compatible servers

- Implemented a new IPC handler in main process to probe for OpenAI-compatible servers and retrieve available models.
- Added a method in preload script to expose backend probing to the renderer.
- Enhanced LLMServerPanel to display server information and allow users to change backend settings.
- Updated settings page descriptions for clarity on server configuration.
- Introduced utility functions to normalize URLs and describe endpoints based on port.

Select highest compatible CUDA assets in build workflow

chore: bump version to 1.0.2 in package.json

refactor: improve RTC connection handling with abort controller and enhanced error management

- implemented stop requests

- improved broken connection recovering logic

Co-authored-by: Copilot <copilot@github.com>

fix: update GitHub repository URL in Quick Start instructions

fixed bug where local server url setting was not being read

Co-authored-by: Copilot <copilot@github.com>

refactored code to use single loop instead of polling mutexes

Co-authored-by: Copilot <copilot@github.com>

fix: update link to Releases page in ReadMe for accuracy

fix: update artifact paths from output to dist in build workflow

fix: prevent publishing during Electron app build