v4.1.1

[4.1.1] - 2026-02-10

Improvements

• Performance: Implemented lazy loading for tabs in Settings and Prompt Editor windows, rendering content only upon selection to improve window opening speed.

Full Changelog: v4.1.0...v4.1.1

v4.1.0

[4.1.0] - 2026-02-09

New Features

• Audio Attachments: Added support for attaching and playing audio files (MP3, WAV, OGG) directly in the chat window.
• Markdown Rendering: Enhanced markdown parsing with support for italicized headers and improved table handling.

Improvements

• Workspace: Improved robustness of frozen state detection to better handle Nuitka and standalone executable environments.

Build & Deployment

• Launchers: Replaced C# launchers with Python-based cx_Freeze implementations to eliminate AV false positives and to simplify and centralize logic.

Full Changelog: v4.0.1...v4.1.0

v4.0.1

[4.0.1] - 2026-02-08

New Features/Changes

• Terminal: Added a new keyboard shortcut [A] to directly open the Audio Analyzer from the console main menu.
• System Tray: Direct file editing options ("Edit config.ini", "Edit prompts.json") in the tray menu are now conditionally visible only when the application is started with the --show-console flag.

Improvements

• GUI: Fix scrolling of Settings and Prompt Editor windows in Tkinter fallback mode.
• UI/UX: Adjusted widget widths and layout in the Settings window for better visual consistency, specifically targeting host/port fields, provider selection, and model dropdowns.
• Terminal: Reorganized the main command box to prioritize the Audio Tool shortcut rather than Endpoint.

Internal

• Request Pipeline: Simplified request origin tracking by consolidating multiple specific endpoint origins into a single ENDPOINT type.

Full Changelog: v4.0.0...v4.0.1

v4.0.0

Changelog

[4.0.0] - 2026-02-08

New Features

Audio Tool

• Audio Analyzer: New tool for recording and analyzing audio from microphone or system loopback.
• Unified Stream Architecture: Robust audio handling for simultaneous recording and level monitoring.
• Visualization: Real-time audio level meter with gradient visualization and resizable UI.
• Controls: Custom input field, action indicator, and response mode toggle (Default, Result Panel, Chat Window).
• Processing: Support for large file recordings and automatic audio compression.
• Tkinter Fallback: Full support for standard Tkinter environments.

System Tray & Navigation

• Added direct access to Direct Chat, Screen Snip, and Audio Analyzer from the tray menu.
• Implemented custom tray menu with separators for better organization.
• Restricted file editing options to console mode.

Settings & Configuration

• Auto-Save: New auto_save_session setting to control when sessions are persisted ("on_followup", "always_window", "on_attachment").
• Reset Defaults: Added button to reset configuration to default values.
• Logging: Centralized logging configuration with Rich formatting support.

Prompts & Actions

• New Actions: Added "Extract Data", "Code Review", "Smart Cleanup", "OCR to Markdown", and more.
• Groups: Support for enabling/disabling action groups in popup menus.
• Overhaul: Comprehensive update to Screen Snip prompts for better results.

Improvements

• UI/UX:
  • Updated blockquote and overlay colors for improved readability across themes.
  • Automatically collapse thinking blocks after streaming finishes.
  • Enhanced Prompt Editor layout and responsiveness.
  • Improved Toast Notification animation and timeout.
  • Button bar visibility improvements at small window sizes.
• Performance:
  • Optimized audio loopback buffer size for better responsiveness.
  • Capture screen snips from frozen background to remove capture delay.

Fixes

• Gemini Provider:
  • Treated "thinking-only" responses as empty to trigger automatic retries.
  • Added detection for invalid API keys in error handling.
• Session Management:
  • Ensured assistant messages are persisted correctly during auto-save.
  • Fixed issue where attachments could be duplicated in API requests.
  • Prevented writing None values to configuration files.
• General:
  • Fixed import paths in prompt and settings windows.
  • Corrected ErrorPopup button styling.
  • Resolved custom input handling issues in Audio Tool.

Refactoring

• Architecture:
  • Centralized multimodal message construction logic.
  • Organized GUI window modules into a dedicated package.
  • Moved attachment handling from session-level to message-level.
  • Removed legacy audio recording implementation.
• Cleanup:
  • Removed AVIF image format support.
  • Converted configuration flags to proper booleans.

Build & Deployment

• Nuitka Migration: Replaced cx_Freeze with Nuitka for standalone executable generation.
• C# Launchers: Introduced lightweight C# launchers to replace Python/Nuitka wrappers, significantly reducing binary size.
• CI/CD: Updated GitHub Actions workflows for the new build process.

Full Changelog: v3.2.0...v4.0.0

Release v3.2.0

🔀 New Feature: Image Comparison

• Compare Images: Select the "Compare Images" action in the Snip Tool popup to start a comparison workflow.
• Two-Step Capture:
  1. Capture your first region (Reference Image).
  2. Capture your second region (Target Image).
  3. The AI will analyze both images together based on your prompt (e.g., "Find differences", "Compare layouts").
• Versatile: Works with standard analysis prompts or custom questions. Just pick a Compare-enabled action or ask a custom question after enabling Compare mode.

🖼️ Enhanced Snip Popup

• UI Update: The popup now supports multi-step interactions for comparison workflows.
• Visual Feedback: Clear indicators show when you're in "Compare Mode" and need to capture a second image.

Full Changelog: v3.1.1...v3.2.0

Release v3.1.1

• Add custom instruction injection as an optional additional step in File Processor tool
• Improve button color for text legibility

Full Changelog: v3.0.0...v3.1.1

Release v3.0.0

Release Highlights v3.0.0

Native Screen Snipping & Major Architecture Update

🆕 New Features

• Built-in Snip Tool (Ctrl+Shift+X): Capture screen regions natively without external tools like ShareX. Includes OCR, image description, and code analysis actions.
• Image Support in Chat: Chat windows now support attaching images and pdf directly to conversations.
• Unified Prompt Configuration: All prompts (text editing, snipping, endpoints) are now centralized in prompts.json for easier management.

⚡ Improvements

• Attachment Manager: Images are now stored as efficient external files (.webp by default) rather than bloating session files, improving load times and memory usage.
• Optimized Prompt Order: Reordered multimodal requests (Media First) for better context adherence with Gemini models.
• Modular GUI: Refactored window management for better stability and maintainability.

⚠️ Breaking Changes

• Flask Server Disabled by Default: The API server (used for ShareX) is now disabled out-of-the-box to reduce overhead. Enable flask_endpoints_enabled = true in config.ini if you need custom integrations.
• Config Migration: text_edit_tool_options.json has been replaced by prompts.json. Old configurations will need to be migrated manually if customized.

Full Changelog: v2.7.12...v3.0.0

v2.6.9

🌟 New Features

• Translation Tools: Added dedicated options to translate text.
• "Direct" Modifier: New 🎯 toggle to make AI responses brief and straight to the point.
• API Key Naming: You can now add descriptive names to your API keys in the settings (e.g., key # Work Account).
• Detailed Model Info: Added context limits, pricing, and "thinking" capability details to the model list in the console.

🖥️ Interface Improvements

• Faster Chatting: Chat windows now open immediately and stream AI responses in real-time.
• Searchable Selectors: Added searchable dropdown menus to easily find AI models in long lists.
• Manual Reordering: Added arrow buttons to customize the order of your favorite actions and modifiers in the Prompt Editor.
• Streamlined Design: Shrink-wrapped the Settings and Prompt Editor windows to take up less screen space.
• Clearer Error Messages: Added simple popup windows to explain exactly what happened if an AI request fails.
• Better Iconography: Refined icons in the terminal and menus for better clarity.

⚙️ Fixes & Optimizations

• Launch Reliability: Fixed an issue where the app would fail to start when opened via Windows Search.
• Background Cleanup: Improved startup speed by moving temporary file cleanup to a background process.
• Windows Terminal Support: Enhanced the show/hide behavior for users running the app in Windows Terminal.
• API Compatibility: Optimized how images and text are sent to AI providers for more reliable results.

Full Changelog: v2.5.4...v2.6.9

v2.5.4

Now supports both legacy UI and Modern UI.
Switch to legacy UI in the Theme tab if experiencing performance issues.

Full Changelog: v2.4.1...v2.5.4

v2.4.1

Download the ModernUI version for better look and latest features.
Grab the older version (AI-Bridge_v2.1.8.zip) below if experiencing performance issues (similar features)

Full Changelog: v2.1.8-ctk...v2.4.1