Skip to content

Releases: iotserver24/codex

CodeX v1.2.2 (Xibe AI)

28 Sep 16:56

Choose a tag to compare

Pre-release

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 Impr...

Read more

v1.1.0

15 Sep 01:02

Choose a tag to compare

v1.1.0 Pre-release
Pre-release

CodeX v1.1.0

TL;DR (What’s new)

  • addition of 2 more providers
  • added prompt library to sae 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.

v1.0.8

28 Aug 18:34

Choose a tag to compare

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.

multiple integration and updates

17 Aug 17:23

Choose a tag to compare

CodeX v1.0.6

🚀 What's New

✨ Major Features

🎨 Designer Mode (Beta)

  • New design-focused chat mode for UI/UX development
  • Optimized for creating beautiful, responsive interfaces
  • Pick any AI model for design tasks
  • Enhanced prompts for visual design and component creation

🤖 Agentic Mode (Coming Soon)

  • Autonomous AI agent for complete project development
  • Structured control loop: Plan → Act → Observe → Reflect → Repair
  • Batch request handling for multiple tasks
  • Context-aware iterative execution

🧠 Thinking Budget System

  • Configurable AI reasoning depth with three levels:
    • Low: Minimal thinking tokens for faster responses and lower costs
    • Medium: Balanced thinking for most conversations (default)
    • High: Extended thinking for complex problems requiring deep analysis
  • Optimized for cost-effective AI interactions

🔧 Auto-Fix Problems

  • Automatic detection and fixing of TypeScript errors
  • Intelligent problem resolution with up to 2 retry attempts
  • Real-time error checking and correction
  • Seamless integration with the development workflow

Native Git Support

  • Optional native Git integration for improved performance
  • Faster Git operations when Git is installed locally
  • Fallback to built-in Git implementation when native Git is unavailable
  • Enhanced Git workflow with better error handling

🔄 Auto-Update System

  • Automatic application updates with configurable release channels
  • Stable Channel: Recommended for most users with reliable updates
  • Beta Channel: Early access to new features with more frequent updates
  • Smart update notifications and download management

🆓 Free AI Models Expansion

New Free Models Added:

  • Claude 3.5 Haiku - Fast and efficient reasoning
  • Gemini 2.5 Flash Lite - Quick responses with good quality

Total: 27+ AI models including premium and free options for every use case!

🛠️ Enhanced Development Tools

Improved Chat Modes:

  • Build Mode: Generate and edit code with AI assistance
  • Ask Mode: Get answers and explanations about your codebase
  • Designer Mode: Focus on UI/UX design and component creation
  • Agentic Mode: Autonomous project development (coming soon)

Better Problem Management:

  • Real-time TypeScript error detection
  • Automatic problem fixing with intelligent retry logic
  • Visual problem reporting with file and line information
  • Manual "Fix All" button for bulk error resolution

Enhanced Settings & Configuration:

  • Thinking budget controls for AI reasoning depth
  • Release channel selection (Stable/Beta)
  • Auto-update preferences
  • Native Git integration toggle
  • Auto-fix problems configuration

🔒 Security & Stability

Improved Error Handling:

  • Better error recovery and user feedback
  • Graceful fallbacks for failed operations
  • Enhanced logging and debugging capabilities
  • More robust file system operations

Performance Optimizations:

  • Faster Git operations with native support
  • Optimized AI model loading and switching
  • Improved memory management
  • Better resource utilization

📱 User Experience Improvements

Enhanced UI/UX:

  • Improved chat mode selector with clear descriptions
  • Better settings organization and accessibility
  • Enhanced notification system
  • More intuitive update management

Better Feedback:

  • Clear progress indicators for long operations
  • Improved error messages and suggestions
  • Better toast notifications for user actions
  • Enhanced help and documentation

📦 Downloads

Windows:

  • Windows (x64): Download the .exe installer for 64-bit Windows
  • Windows (ARM64): Download the .exe installer for ARM64 Windows

macOS:

  • Mac Universal: Download the universal .dmg file (works on both Intel and Apple Silicon)
  • Mac (ARM64): Download the .dmg file optimized for Apple Silicon Macs
  • Mac (x64): Download the .dmg file for Intel Macs

Linux:

  • Linux (x64) DEB: Download the .deb package for Debian/Ubuntu 64-bit systems
  • Linux (ARM64) DEB: Download the .deb package for Debian/Ubuntu ARM64 systems
  • Linux (x64) RPM: Download the .rpm package for Red Hat/Fedora 64-bit systems
  • Linux (ARM64) RPM: Download the .rpm package for Red Hat/Fedora ARM64 systems

🔧 Installation

  1. Download the appropriate installer for your platform and architecture
  2. Run the installer and follow the setup wizard
  3. Launch CodeX and start coding with AI!

⚠️ Note

These builds are not code-signed and may show security warnings. This is normal for open-source applications.

🐛 Bug Reports

If you encounter any issues, please report them on our GitHub Issues page.

📚 Documentation

Visit our documentation for detailed guides and tutorials.

🔄 Updating from v1.0.4

If you're updating from v1.0.4, you'll notice:

  • New chat modes in the interface
  • Additional free AI models available
  • Enhanced settings with new configuration options
  • Improved error handling and problem resolution
  • Better performance with native Git support

🎯 What's Next

  • Agentic Mode: Full autonomous AI development capabilities
  • Enhanced Designer Mode: More specialized design tools
  • Additional AI Models: More free and premium model options
  • Advanced Git Integration: Enhanced version control features

Built with ❤️ by the CodeX team

vercel V1.0.4 vercel integration

13 Aug 01:22

Choose a tag to compare

CodeX v1.0.4

🚀 What's New

  • AI-powered code generation and editing
  • 40+ AI models support including GPT-5, Claude, and free models
  • Vercel integration for one-click deployment
  • Built-in terminal and development tools
  • Cross-platform support (Windows, macOS, Linux)

📦 Downloads

Windows:

  • Installer: Download the .exe file for easy installation
  • Portable: Download the .zip file for portable use

macOS:

  • Download the .zip file and extract to Applications

Linux:

  • Debian/Ubuntu: Use the .deb package
  • Red Hat/Fedora: Use the .rpm package
  • Other: Use the .zip file

🔧 Installation

  1. Download the appropriate file for your platform
  2. Run the installer or extract the zip file
  3. Launch CodeX and start coding with AI!

⚠️ Note

These builds are not code-signed and may show security warnings. This is normal for open-source applications.

🐛 Bug Reports

If you encounter any issues, please report them on our GitHub Issues page.

📚 Documentation

Visit our documentation for detailed guides and tutorials.


Built with ❤️ by the CodeX team

version: 1.0.2

17 Jul 14:06
9523100

Choose a tag to compare

🎉 Release 1.0.2 🎉

Hello everyone! We're thrilled to bring you version 1.0.2. This update is packed with new features and important fixes to make your experience even better.


✨ What's New?

  • Check for Updates: Now you can easily check for the latest version directly from within the application. Stay up-to-date with a single click!
  • Support Our Work: We've added a "Donate" button. If you love our project, please consider supporting its future development. Every little bit helps!

🐛 Bugs Squashed!

  • We've fixed several minor bugs to improve stability and performance, making the app more reliable for you.

📝 Report an Issue

Found a bug? Have a suggestion? We'd love to hear from you! You can create an issue by clicking the link below:

Click here to open a new issue


A huge thank you to our amazing community for your continued support and feedback!

Fixed api calls and more

13 Jul 10:15
65f5f18

Choose a tag to compare

In this update ,many issues has been fixed and given new updates

Release v1.0.0

11 Jul 09:00
b6c15e7

Choose a tag to compare

added logo