Changelog

Messaging credential input fields now expand to fill their available space, making it easier to see and enter information like bot tokens. This fixes an issue where placeholders were cut off.

The switch for enabling messaging platforms now clearly shows its current status as "Enabled" or "Disabled" next to it. This makes it much easier for users to understand what the switch controls and its current setting.

When no scheduled jobs exist for a profile, a "New cron" button now appears in the detail pane, making it easier to create your first scheduled job. This fixes an issue where it was harder to start creating scheduled jobs if you had no existing ones.

When the agent fails to start, you'll now only see one error message instead of two. This makes it clearer when Hermes encounters a problem.

The application now shows a confirmation message when you successfully apply a new main model. This makes it clear that your chosen model has been saved and will be used for future chats.

Assistant replies now correctly preserve single line breaks, so things like poems, addresses, or lists that are meant to be displayed line by line will appear as intended. This fixes an issue where these lines were incorrectly merged into a single paragraph.

The desktop application now displays helpful messages when it's waiting for the AI provider to respond, letting you know why it's delayed and how long it might take. This makes it clearer when the provider is temporarily unavailable or has been limited, and shows how to stop the wait if needed.

Error messages now show the friendly name of the service that caused the problem, rather than its internal code name. This makes it easier to understand which service is having an issue.

The error message when the AI service cannot be reached has been improved to be more accurate. It now clearly states if the service couldn't be reached at all or if it didn't respond in time, providing better guidance for users.

Custom models you save are now correctly recognized in the settings and bot model selection menus. This ensures your saved custom models appear as expected, instead of showing as unconfigured.