This pull request introduces a major new feature: user-selectable Copilot models for alt text generation, along with improvements to documentation and settings management. Users can now choose from available vision-capable models via a dropdown in the settings, with the list dynamically fetched from the Copilot SDK. The default model is now configurable, and all relevant documentation has been updated.

Model Selection and Settings Integration

• Added a settings dropdown in SettingsTab.jsx to select the Copilot model for alt text generation, with available models fetched dynamically and filtered to only show vision-capable options. The selected model is persisted in settings and used for all alt text generation jobs. [1] [2] [3] [4] [5] [6]

Copilot SDK Integration

• Implemented the listModels() API in copilot-adapter.js to query available models from the Copilot SDK, detect vision capability, and provide model metadata. Added IPC and preload wiring to expose this API to the renderer. [1] [2] [3] [4]

Documentation Updates

• Updated the README.md to include installation instructions for pre-built binaries, describe the new model selection feature in the settings table, and clarify the build process. [1] [2]
• Added a new changelog entry for version 0.3.0, describing the model selection and SDK integration features.

Version Bump

• Updated the application version to 0.3.0 in package.json to reflect the new features.