1、Cloud conversation sync across multiple devices 2、First-use onboarding guide for quick server setup 3、MCP tool call card and command output highlighting 4、Chat favorites with pinned section display 5、Upgraded to Inter font for refined visual experience

Fix the abnormal detection issue of the new version

1. Add context window management and intelligent auto title generation for AI conversations 2. Optimize update dialog UI and Android app store redirect links 3. Improve SSH connection stability with tmux support

1.Bastion host connection reuse 2.Add a "Recent Connections" group to the quick connection list 3.Fix other model interaction issues

1.Fix the problem of probabilistic duplicate execution in the model 2.Fix compatibility issues with data synchronization for low-end models

What's New Interactive commands — Run commands like sudo that need password input Asset sync — Keep your assets in sync across devices Device management — Manage trusted devices in one place Task completion alerts — Get notified when an AI task finishes Quick start from empty state — See common tasks and start a conversation with one tap Custom rules — Set your preferred tone and style for AI replies Improvements Command execution — Automatic escaping before run to avoid failures and edge cases Tool messages — Clearer, easier-to-scan context tool messages Bug Fixes Fixed task being interrupted when the model replied without using tools Fixed execution failures when single quotes inside the command conflicted with outer quoting (e.g. free | ...) Fixed data synchronization issues in some versions when upgrading to 1.1.0.

What's New Interactive commands — Run commands like sudo that need password input Asset sync — Keep your assets in sync across devices Device management — Manage trusted devices in one place Task completion alerts — Get notified when an AI task finishes Quick start from empty state — See common tasks and start a conversation with one tap Custom rules — Set your preferred tone and style for AI replies Improvements Command execution — Automatic escaping before run to avoid failures and edge cases Tool messages — Clearer, easier-to-scan context tool messages Bug Fixes Fixed task being interrupted when the model replied without using tools Fixed execution failures when single quotes inside the command conflicted with outer quoting (e.g. free | ...) Fixed extra blank line below Markdown content

Address some known issues and optimize the user experience -------Improvements------- 1. Streaming speech recognition in terminal command mode 2. One-click AI command interpretation 3. Auto-execute confirmation button moved to command message bottom

add user guide in Profile -> Help & Feedback

Optimize multiple details to deliver a smoother user experience

update model support