🔧 Bug Fixes

• Category tree editing (Hub + CPB): Single-click filter is deferred so double-click rename no longer refreshes the list before the inline editor opens.
• Custom Prompt Builder — prompt rows to categories: Grid.js re-renders no longer break drag-and-drop; rows stay in sync via observers and delegated dragstart.
• Category drag-and-drop: Reorder/reparent rules for roots and subcategories are completed; dropping on self or Uncategorized is a safe no-op.

⬆️ Improvements

• Drop highlight & rename focus: Visual highlight on category targets while dragging prompts; inline rename field regains focus after tree rebuilds when appropriate.

✨ New Features

• Left Panel: Dashboard + Prompt Hub: Added two dedicated entries in the left rail so users can jump directly into dashboard insights and prompt workflows.
• Custom Prompt category management: Added category management flow for root/sub categories and prompt assignment in the Custom Prompt Builder and Prompt Hub workflow.

⬆️ Improvements

• Chat History visibility: Improved left rail visual clarity so the Chat History menu is easier to spot and access.
• Mermaid live preview: Improved Mermaid block render stability with better fullscreen/fit behavior and more reliable preview redraw.

⬆️ Improvements

• Gemini idle refresh preserves conversation: The Gemini webview now reloads the current conversation URL in-place instead of navigating to a new chat. Your multi-panel conversation context is no longer lost after each refresh cycle.
• Typing-aware refresh pause: While typing in the main prompt input, the Gemini refresh timer automatically pauses (blinking countdown) and resumes 3 seconds after you stop typing—no more focus stealing mid-composition.
• Scroll & caret restoration: After an idle refresh, the Gemini webview's scroll position on conversation URLs is saved and restored. The main prompt input's caret position and text selection are also preserved.

⬆️ Improvements

• Perplexity login detection: Clearer logged-in vs logged-out detection (KO/EN CTAs, auth paths, visibility-based login UI checks, short grace period). Expanded login selectors in config. The Login Required badge should track real session state more reliably.
• External login via Microsoft Edge: If Google Chrome is not installed, external browser login uses Microsoft Edge (Chromium) on Windows and macOS—Edge-only environments can complete Gemini and other Google-based cookie sync without installing Chrome.

✨ New Features

• Prompt Builder: Mermaid, Code, LaTeX Preview: In the Custom Prompt Builder and main prompt Live Preview, code blocks (with syntax highlighting), Mermaid diagrams, and LaTeX formulas are now rendered in dedicated preview sections. Mermaid supports zoom, fit, and fullscreen; LaTeX is rendered with KaTeX.

⬆️ Improvements

• Gemini session extension period relaxed: Keep-alive and idle-refresh intervals are tuned so the Gemini session is less likely to expire during idle. A visible countdown timer per Gemini panel shows when the session will be extended.

🔧 Bug Fixes

• Preview variable rendering: Fixed {{last_response}} and {{chat_thread}} appearing truncated in the main prompt inline and master preview.
• Anonymous mode and variable substitution: When Anonymous is ON, variables sent to webviews now use anonymized service names; all previews update in real time.
• Expand prompt area after send: Sending a prompt (Send or Ctrl+Enter) now automatically collapses the expanded prompt area so you can see AI responses immediately.

⬆️ Improvements

• Gemini session keep-alive: A 5-minute keep-alive helps reduce Gemini logout when the app is idle for ~10+ minutes.

✨ New Features

• Custom Prompt Builder: Full-screen 3-panel editor with System/Global/Local variables, autocomplete, Send and Copy to Input.
• Slash Command: Type / in the main input to insert saved custom prompts with keyboard navigation.
• Inline Variable Form: Auto-generated mini form for {{variable}} tokens; unresolved variable confirmation before send.

⬆️ Improvements

• Popup Layer: Slash list and variable UI appear above webviews with dim/blur; no BrowserView resize.
• File Upload Modal: No longer hidden behind webviews; countdown and progress bar UX; English text.
• Gemini One-Login: One Chrome sign-in can reflect in webview; first sync on service domain, re-sync when cookie count increases.
• Perplexity: Faster file+prompt flow; reduced duplicate prompt injection.

🔧 Fixes

• File Upload Modal Timer: Duplicate timer prevented; layer state restored correctly when another popup is active.
• Preview + Collapse: Collapse controls now work when Live Preview is ON.

🔧 Fixes

• Google CookieMismatch Fix: Resolved cookie mismatch error that prevented Google login for Gemini and Genspark services.
• Third-Party Cookie Policy: Disabled Chromium 134+ third-party cookie deprecation blocking Google's cross-domain auth flow.
• Dynamic User-Agent: Replaced hardcoded Chrome/130 with actual Chromium version to prevent bot detection.

✨ New Features

• Single AI Mode: For users who subscribe to only one AI service—open up to 3 instances of the same provider and compare responses from different models.
• Anti-Bot Detection: Staggered timing for instance creation and prompt delivery to reduce bot detection. Human-like typing simulation included.
• Chat Mode Settings: New gear icon (⚙️) and modal UI for switching between Multi AI Mode and Single AI Mode.

🔧 Fixes

• Memory Leak Fix: BrowserView instances are now properly cleaned up when switching modes.
• URL Persistence: Improved conversation URL tracking for Single Mode instances.

🔧 Fixes & Improvements

• Gemini/Genspark Login Fix: Fixed login not persisting after external Chrome login when user had logged out within the app.
• Update Download Progress: Added visual download progress window with progress bar during update download.
• Taskbar Progress: Windows taskbar now shows download progress indicator.

🔧 Fixes

• Update Dialog: Fixed HTML tags being displayed in update dialog release notes. Release notes are now shown as plain text.

🔧 Fixes & Improvements

• Gemini Login Persistence: Fixed Gemini login not persisting on app restart by syncing accounts.google.com cookies and improving cookie name detection.
• Reload Button Repositioned: Moved reload button from URL bar to header (left of maximize button) for better accessibility.
• Per-Service New Chat: Added "New Chat" button (plus icon) in URL bar for each service panel, allowing individual service chat resets.

✨ New Features

• History Bulk Delete: Select multiple sessions in the History Sidebar and delete them in one action (Selection Mode).
• History UX Improvements: Improved infinite scroll and scroll restoration for long history lists.
• Modernized Selection UI: Refined selection toolbar styling and interactions for better clarity.

✨ New Features

• Custom Prompt Persistence: Saved prompts now sync to the app data store (electron-store) with localStorage migration to prevent loss on reinstall.
• Login UX Fix: “Chrome으로 로그인” badge reset fixed (ID mismatch) so the button clears after closing the Chrome login window.
• Cookie Sync Hardening: SameSite mapping and default path fixes to keep ChatGPT/Claude/Gemini/Grok sessions stable on Windows 11.

✨ New Features

• Installer Distribution: Now distributed as a Windows installer (.exe) for easier installation.
• Auto-Update: Automatic update support via electron-updater. Get the latest features without manual downloads.
• GitHub Actions CI/CD: Automated build and release pipeline for faster deployments.

🐛 Bug Fixes

• Grok Login Fix: Resolved "Login required" issue by syncing Google/X cookies.
• Login Detection: Improved state detection for Grok.

✨ New Features

• Genspark Integration: Full support for search-driven AI chat.
• Anonymous Mapping: Maps Genspark to alias "(F)".

✨ Conversation History

• New collapsible sidebar for easy history access.
• Auto-Save functionality for sessions, layouts, and URLs.
• Optimized asynchronous IndexedDB operations.

✨ Session Persistence

• Saves active services, layout, scroll sync, and URLs on close.
• WebView URL Bar: Dedicated URL bar with Reload, Copy, and Open in Browser buttons.
• 3x1 Vertical Layout: New layout option for vertically stacked services.

✨ Enhanced Copy

• Extracts full conversation with role distinction (User/AI).
• Copy Last Response: Button to copy only the last AI response.
• Per-Service Header: Dedicated header for each service panel.

✨ Cross Check Enhancements

• Editable predefined prompts and custom prompt management.
• Robust Input State: Fixed issues with input fields becoming disabled.

✨ Anonymous Cross Check

• Replace service names with aliases.
• Modern UI: Scroll sync toggle, layout icons, and refined styling.

✨ File Upload Support

• Multi-service broadcasting for files and images.
• Drag & Drop: Direct file attachment support.