The last version update post on this forum was AgentOne Desktop v0.0.18 is out now!. Below are the changes between versions 0.0.20 and 0.0.21 of the desktop app. To view the changes between 0.0.18 and the current version, open the desktop app settings, click Help & Updates, and expand the release notes explorer.
Highlights
- New: You can now customize keyboard shortcuts in settings.
- New: AgentOne can now store memories across chats! They sync with your account when settings sync is enabled and are passed to the AI as context for each chat.
- Notice: We are removing the multi-window feature. AgentOne is currently limited to a single window. Please discuss this change on the forum.
- Bug Fix: Reduced the memory usage of the application. More memory usage fixes are coming - stay tuned!
- Improved the Account section of the settings page.
- Added nearly 1000 new extensions.
- Now, you can view release notes from previous versions (and the current version) at any time in
Settings > Help & Updates. - New: You can now configure text-to-speech (TTS), which lets AgentOne read messages aloud. You can select a provider and voice in the
Providers > Text-to-Speechsection of settings.
Miscellaneous Changes
- Clicking a chat in the sidebar that’s open in another window no longer loads the chat by default in the background while the dialog is displayed in the foreground.
- Fixed a bug where notifications wouldn’t send on Linux. Due to a bug in a notification service we use, notifications still may not function on Wayland.
- Tweaked the splash screen (again
) - Using the desktop app right after creating your account will alert you that your billing details may still be setting up. Previously, the app would fail to send messages with no explanation.
- Updated various app components to improve performance.
- Keyboard shortcuts are now triggered even when you are typing in an input field.
- Fixed a bug in our analytics tracking.
- Adjusted the contrast of the sidebars in the application.
- Fixed the alignment of the placeholder in the chat model selector.
- You can now customize the maximum output tokens for AI title generation in chat settings.