CodeX v1.2.2 (Xibe AI)

CodeX v1.2.2 (Xibe AI)

TL;DR (What's new)

• Complete app rebrand - Now called "Xibe AI" with new domain xibe.app
• MCP Tools Integration - Connect and use external tools via Model Context Protocol
• Enhanced Agent-MCP Mode - Advanced AI agent with MCP capabilities
• Share Preview Option - Preview shared apps before publishing
• Universal Thinking Support - All AI models now support thinking capabilities
• E2B Integration - Share app previews with secure sandboxes
• Free Pro Features - All premium features now available at no cost

Highlights

🎯 Complete App Rebrand - Xibe AI

• New Identity - Renamed from "CodeX" to "Xibe AI" for better brand recognition
• New Domain - Migrated from codex.anishkumar.tech to https://xibe.app
• Updated Branding - New logos, icons, and visual identity throughout the app
• Consistent Naming - All references updated across the application

🔧 MCP Tools Integration

• External Tool Connections - Connect to external tools and services via MCP
• Seamless Integration - Use connected tools directly within AI conversations
• Tool Management - Easy-to-use interface for managing MCP tool connections
• Enhanced Capabilities - Expand AI functionality with external integrations

🤖 Enhanced Agent-MCP Mode

• Advanced AI Agent - Intelligent agent with MCP protocol support
• Tool Orchestration - Agent can automatically use connected MCP tools
• Smart Context - Better understanding and utilization of available tools
• Autonomous Operations - Agent can perform complex multi-step tasks

🌐 Share Preview Option

• Preview Before Sharing - Test shared apps before making them public
• Real-time Testing - See exactly how shared apps will appear to others
• Instant Feedback - Quick validation of shared app functionality
• Secure Preview - Safe testing environment before public sharing

🧠 Universal Thinking Support

• All AI models now support thinking - No more division between thinking and non-thinking models
• User-controlled thinking intensity - Choose Low/Medium/High budget settings for any model
• Consistent thinking experience - Same advanced reasoning capabilities across all providers
• Provider-specific optimizations - Tailored thinking configurations for different AI providers

🌐 E2B Integration & Sharing

• E2B Sandbox Sharing - Share your app previews publicly via E2B sandboxes
• Duration Controls - Choose from 5/10/15 minutes for sharing
• Real-time Status - Live monitoring of sandbox status, logs, and progress
• Automatic URL Copying - Share links are automatically copied to clipboard

🐛 Bug Fixes (v1.1.2)

• Fixed getCodexAppPath error - Resolved ReferenceError: getCodexAppPath is not defined that was causing chat functionality to fail
• Fixed codexApiKey undefined error - Resolved missing API key configuration for CodeX Pro features
• Fixed database migration issues - Resolved conflicts in database schema and migration files
• Fixed React rendering errors - Resolved missing imports and undefined references in React components
• Fixed MCP table creation - Resolved missing mcp_servers and mcp_tool_consents tables in database

🔧 Technical Improvements

Enhanced AI Integration

• Universal Thinking - All AI models now support advanced reasoning and thinking capabilities
• Better Performance - Improved response quality and speed across all providers
• Smarter Responses - Models can now think through complex problems before responding
• User Control - Adjust thinking intensity (Low/Medium/High) for any model

MCP Protocol Support

• Seamless Tool Integration - Connect external tools via standardized MCP protocol
• Enhanced Agent Capabilities - Agents can now utilize connected tools autonomously
• Better Context Management - Improved handling of tool capabilities and context

Sharing & Collaboration

• Preview Functionality - Test shared apps before publishing
• Public App Sharing - Share app previews with anyone via secure sandboxes
• Real-time Monitoring - Watch shared apps in real-time with live status updates
• Flexible Duration - Choose appropriate sandbox duration for sharing needs

🆓 Free Features

• CodeX Pro is now free - All Pro features including web search, smart context, and turbo edits are now available at no cost
• Agent-MCP mode - Renamed "Agent" mode to "Agent-MCP" to better reflect MCP (Model Context Protocol) integration
• Enhanced MCP support - Improved Model Context Protocol integration for better tool usage
• Universal Thinking - Advanced reasoning capabilities across all AI models
• App Sharing - Share creations with secure, temporary sharing links

📚 Documentation Updates

• Updated all documentation to reflect v1.2.2 changes
• Added comprehensive MCP tools integration guide
• Updated branding references from CodeX to Xibe AI
• Enhanced sharing and preview documentation
• Fixed version references in getting started guides
• Updated troubleshooting documentation

🚀 Performance Improvements

• Faster startup - Reduced application startup time
• Better memory management - Improved memory usage and garbage collection
• Optimized database queries - Faster database operations
• Enhanced tool loading - Quicker MCP tool connection and initialization
• Improved sharing performance - Faster sandbox creation and sharing

🎨 UI/UX Improvements

• Complete rebrand - New Xibe AI visual identity throughout the application
• Enhanced settings panel - Cleaner, more intuitive settings interface
• Better chat experience - Improved chat UI and message handling
• Improved tool management - Better interface for MCP tool connections
• Enhanced app management - Better app creation and management workflow
• Redesigned sharing interface - Comprehensive sharing controls with preview options

New Features & Capabilities

• MCP Tools Connection - Connect to external tools and services
• Agent-MCP Mode - Advanced AI agent with tool orchestration
• Share Preview - Test shared apps before publishing
• App Rebranding - Complete transition to Xibe AI identity
• Enhanced Sharing - Improved sharing workflow with preview capabilities
• Universal Thinking - Advanced reasoning across all AI models

Breaking Changes

• Brand Name Change - Application name changed from "CodeX" to "Xibe AI"
• Domain Migration - Website moved from codex.anishkumar.tech to https://xibe.app
• None - All other changes are backward compatible

What's New for Users

• New App Identity - Welcome to Xibe AI, the evolution of CodeX
• External Tool Integration - Connect and use external tools seamlessly
• Advanced AI Agents - Experience intelligent agents with MCP capabilities
• Preview Sharing - Test your shared apps before making them public
• Universal AI Thinking - All models now support advanced reasoning
• Enhanced Sharing - Better app sharing with preview and duration controls
• Free Pro Features - All premium features remain free to use

Performance & Reliability

• Faster Responses - Optimized AI model performance across all providers
• Better Stability - Improved error handling and recovery
• Secure Sharing - Shared apps run in isolated, secure environments
• Tool Integration - Reliable MCP tool connections and usage
• Smoother Experience - Reduced loading times and more responsive interface

Summary

• Major Rebrand - Complete transition to Xibe AI identity and domain
• MCP Integration - External tool connections via Model Context Protocol
• Enhanced Agents - Advanced AI agents with tool orchestration
• Preview Sharing - Test shared apps before publishing
• Universal Thinking - Advanced reasoning across all AI models
• Bug Fixes - Comprehensive fixes from unreleased versions
• Performance - Improved speed, stability, and user experience

---

Previous Versions

CodeX v1.2.0

TL;DR (What's new)

• Universal Thinking Support - All AI models now support thinking capabilities
• E2B Integration - Share your app previews with the world via E2B sandboxes
• Enhanced Share Panel - New sharing interface with duration controls
• Improved Model Support - Extended thinking capabilities across all providers
• Better Error Handling - Enhanced error management and user feedback

Highlights

🧠 Universal Thinking Support

• All models now support thinking - No more division between thinking and non-thinking models
• User-controlled thinking intensity - Low/Medium/High budget settings work for all providers
• Provider-specific optimizations - OpenAI gets reasoning_effort + thinking, others get thinking config
• Backward compatibility - Existing thinking models continue to work as before

🌐 E2B Integration & Sharing

• E2B Sandbox Sharing - Share your app previews publicly via E2B sandboxes
• Duration Controls - Choose from 5/10/15 minutes for sharing
• Real-time Status - Live monitoring of sandbox status, logs, and progress
• Automatic URL Copying - Share links are automatically copied to clipboard

🎨 Enhanced Share Panel

• New Share Interface - Dedicated share panel with comprehensive controls
• Duration Selection - Easy dropdown for choosing sandbox duration
• Status Monitoring - Real-time display of sandbox status and logs
• Progress Tracking - Visual progress indicators for sandbox operations
• Version Management - Track and manage multiple sandbox versions

Technical Improvements

Enhanced AI Model Support

• Universal Thinking - All AI models now support advanced reasoning and thinking capabilities
• Better Performance - Improved response quality and speed across all providers
• Smarter Responses - Models can now think through complex problems before responding
• User Control - Adjust thinking intensity (Low/Medium/High) for any model

Sharing & Collaboration

• Public App Sharing - Share your app previews with anyone via secure sandboxes
• Real-time Monitoring - Watch your shared apps in real-time with live status updates
• Flexible Duration - Choose how long your shared apps stay online (5-15 minutes)
• One-Click Sharing - Share links are automatically copied to your clipboard

New Features & Capabilities

• App Sharing - Share your built apps with others via secure online sandboxes
• Thinking Controls - Fine-tune how much AI models think before responding
• Better Error Messages - Clearer feedback when things go wrong
• Improved Settings - Enhanced configuration options for better control

User Experience Improvements

• Intuitive Sharing - One-click sharing with automatic URL copying
• Visual Feedback - Clear status indicators and progress tracking
• Flexible Duration - Choose appropriate sandbox duration for your needs
• Better Error Recovery - Clearer error messages and helpful guidance
• Smoother Performance - Faster app loading and more responsive interface

Breaking Changes

• None - All changes are backward compatible

What's New for Users

• All AI Models Now Think - Every model can now use advanced reasoning, not just specific ones
• Share Your Apps - Show your creations to others with secure, temporary sharing links
• Better AI Responses - Models think more carefully before responding, leading to better results
• Improved Reliability - Fewer errors and clearer feedback when issues occur

Performance & Reliability

• Faster Responses - Optimized AI model performance across all providers
• Better Stability - Improved error handling and recovery
• Secure Sharing - Your shared apps run in isolated, secure environments
• Smoother Experience - Reduced loading times and more responsive interface

Summary

• Major Feature Addition - App sharing capabilities for collaboration and showcasing
• Universal AI Enhancement - All models now support advanced thinking and reasoning
• Better User Experience - Improved performance, reliability, and error handling
• No Breaking Changes - Everything works exactly as before, just better

CodeX v1.1.0

TL;DR (What's new)

• addition of 2 more providers
• added prompt library to save the prompts
• added docs (beta)
• Added 5 new themes with app-wide color variables
• Improved text contrast across the UI (inputs, pills, cards)
• Theme‑aware chat inputs and title bar window controls
• Settings and Hub now use theme tokens (no hardcoded colors)
• Temporarily hid the "Backend Services" card on Hub
• Improved macOS CI: upload correct artifacts for universal builds

Highlights

• Themes
  • New themes: Retro‑Future Tech, Urban Mysteries, Micro‑Adventures, AI + Art Experiments, Digital Magic
  • Each theme defines foreground/on‑color tokens for accessibility
• Readability & Contrast
  • Input boxes, action icons, headings, and helper text now bind to foreground, muted-foreground, bg-card, and border tokens
  • Better visibility on both light and dark backgrounds
• Title Bar
  • Minimize / Maximize / Close icons adopt sidebar theme colors with proper hover behavior
• Hub
  • "Backend Services" section is hidden for now
• CI/CD
  • macOS release workflow collects x64 and arm64 zips generated by the universal step

Fixes & polish

• Removed hardcoded grays in Settings and other screens in favor of theme tokens
• Minor UI consistency tweaks across buttons and labels

Notes

• This version focuses on visual polish and accessibility; functionality remains unchanged.

CodeX v1.0.8

TL;DR (What's new)

• Added 5 new themes with app-wide color variables
• Improved text contrast across the UI (inputs, pills, cards)
• Theme‑aware chat inputs and title bar window controls
• Settings and Hub now use theme tokens (no hardcoded colors)
• Temporarily hid the "Backend Services" card on Hub
• Improved macOS CI: upload correct artifacts for universal builds

Highlights

• Themes
  • New themes: Retro‑Future Tech, Urban Mysteries, Micro‑Adventures, AI + Art Experiments, Digital Magic
  • Each theme defines foreground/on‑color tokens for accessibility
• Readability & Contrast
  • Input boxes, action icons, headings, and helper text now bind to foreground, muted-foreground, bg-card, and border tokens
  • Better visibility on both light and dark backgrounds
• Title Bar
  • Minimize / Maximize / Close icons adopt sidebar theme colors with proper hover behavior
• Hub
  • "Backend Services" section is hidden for now
• CI/CD
  • macOS release workflow collects x64 and arm64 zips generated by the universal step

Fixes & polish

• Removed hardcoded grays in Settings and other screens in favor of theme tokens
• Minor UI consistency tweaks across buttons and labels

Notes

• This version focuses on visual polish and accessibility; functionality remains unchanged.

Full Changelog: v1.1.0...v1.2.2