Versionsverlauf

- Added support for iPhone Duo - Changed the Live Activity text color to make it easier to read - Fixed an issue where the Live Activity in the Dynamic Island could sometimes appear distorted - Made some other adjustments

- Added support for some features available on iOS 27 and later - Increased the minimum OS requirement to iOS 27 - Added Private Cloud Computing as an option for Apple Foundation Models - Private Cloud Computing may be removed in the future - Added reasoning level settings for supported models in Apple Foundation Models - Added the Claude API as a new assistant - You can use supported models with your own API key - Available as a beta version, so there may be bugs - Added sidebar support - Moved the assistant list to the sidebar - Updated the app icon - Dismissed notifications from the Notification Center after using the assistant in the background - Fixed an issue where new chats would no longer be saved if there was too much chat history - Made some other adjustments

- Made Apple Foundation Models available in China mainland when available on device - Changed haptics to occur even when sound effects are disabled - Changed the default setting for sound effects to "Background Only"

- Added an option to use a Bluetooth microphone - Made some other adjustments

- Applied the URL opening option to all assistants - Added buttons to copy and share URLs provided by the assistant in the context menu - Added shortcuts to change the following settings - Assistant - Start with Mic Muted - URL Opening Option - Continue in Background - Keep on Standby in Background - Sound Effects Mode - Made some other adjustments

- Applied the URL opening option to all assistants - Added buttons to copy and share URLs provided by the assistant in the context menu - Added shortcuts to change the following settings - Assistant - Start with Mic Muted - URL Opening Option - Continue in Background - Keep on Standby in Background - Sound Effects Mode - Made some other adjustments

- Made the in-app browser close when starting the assistant - Added an option to disable web search in the Google Gemini API - Fixed an issue where Live Activities would sometimes not end - Made some adjustments to optimize the microphone and speaker for a better experience - Made some other adjustments

- Added sound effects - Added a search bar to the chat history - Enabled resuming the assistant from chat history - Added automatic keyboard focus when "Start with Mic Muted" is enabled - Added close and submit buttons to the keyboard - Removed the "Assistant Screen Size" setting - Made some other adjustments

- Added the assistant status to the Live Activity - Added an option to allow the assistant to remain on standby in the background - By keeping the microphone on, you can have the assistant standby in the background - While in standby, you can use the Side Button or Action Button to resume the assistant without opening the app - Reduced the time until automatic speech recognition ends from 30 seconds to 10 seconds - Enabled extending the time until automatic speech recognition ends by using Side Buttons or Action Button - Added a shortcut to stop the assistant while it is in use - Allowed access to audio content from other apps even while Side Search is using the microphone - Fixed an issue where the app could crash when turning the microphone back on after the input source has been changed - Made some other adjustments

- Enabled continuous speech recognition in the background even without notification permissions - Added a setting to choose whether to continue speech recognition in the background - If possible the assistant screen will be displayed immediately when starting the assistant from the background - Fixed an issue where Live Activities might not start - Fixed bugs - Made some other adjustments

- Restored the default assistant screen size to full screen - Enabled continuous conversation with assistant in the background - To use this, you need to allow notifications - If you start a conversation in the background while using a supported assistant, you will be asked to allow notifications - URL Based Assistant are not supported - If you remain silent for 30 seconds after the assistant replies, the assistant will automatically terminate - Live Activity now starts when conversing with assistant in the background - Prevented system Auto-Lock when using speech recognition - Made some changes to the search engine presets - Added Google AI mode - Changed Gemini's URL - Made some other adjustments

- Made it possible to start the assistant using keyboard shortcuts or VoiceOver Magic Tap - Added support for Assistive Access - Fixed an issue where speech recognition could not be restarted - Fixed an issue where closed modals could be re-displayed - Fixed an issue where multiple windows could be created on the iPad - Made some other adjustments

- Added Markdown support - Added a button to copy a message with the assistant - Added a button to translate a message with the assistant - Added an option to save chat history - Added Pride & Unity icon - Made some other adjustments

- Added assistant screen size setting - Added a control to start the assistant from the Control Center - Fixed an issue where the query part was not correctly replaced in the URL Based Assistant - Fixed an issue where speech recognition could start while the assistant was loading - Made some other adjustments

- Added a setting to switch assistants - Added support for conversational assistants - Added Apple Foundation Models to the assistant - Only available on devices that support Apple Intelligence - Added Google Gemini API to the assistant - You need to obtain your own API Key from Google AI Studio - You can use the web search feature - Redesigned the assistant's UI - Changed the app icon - Made it possible to change the app icon - Changed or removed some settings - Made some other adjustments

- By removing "%s" from the Search URL, you can skip the Side Search's speech recognition - Changed the text "Recommended Search Engines" to "Search Engine Presets" - Removed Grok from the Search Engine Presets - Made some other adjustments

- Enabled URL schemes other than HTTP and HTTPS - Updated help hints - Fixed an issue where sound, including VoiceOver, would not play at all while using speech recognition - Made some other adjustments

- Made some changes to the initial setting of "Speech Language" - Fixed a bug in checking the availability of Speech Recognition - Speech Recognition is processed on-device, so the request has been removed - Made some other adjustments

Release