Updated the app name and icon to microAI, with refreshed localized copy across supported languages. Fixed an issue that could prevent the app from launching on the latest iOS versions, and improved startup reliability and overall stability. Improved connections to Mac and browser compute nodes, including better recovery after temporary network interruptions. Refined parts of the interface to make everyday interactions clearer, more consistent, and easier to use.

**Your computer can now power morai.** With morai 0.3, your AI is no longer limited by your phone. Simply scan a QR code to connect to your computer, then use the AI models running there directly from your phone. Your phone becomes the remote. Your computer does the heavy lifting. This means you can use larger, more capable local AI models while keeping your conversations and data on your own devices. We've also made connections more reliable and improved the overall experience. **Your AI. Your devices. Your data.**

morai 0.2.0 introduces Bees — private AI helpers built for different jobs. • Private Chat — ask, write, brainstorm, and talk privately • Image Chat — understand photos and screenshots • Voice Chat — talk naturally with your local AI • Review Writer — turn real experiences into clear reviews • Photo Tagger — generate titles, descriptions, and keywords for photos Give each Bee the local AI model you want. Use a small model for quick everyday tasks, a vision model for images, or a more capable model when you need it. Faster, lighter local AI We've rebuilt how models run inside morai to make local AI feel smoother and more efficient on your device. Models now load and switch more intelligently, unnecessary reloads are reduced, and memory, thermal, and background resources are managed more carefully. That means a smoother experience across chat, voice, vision, and memory tasks. Easier to get started New to local AI? morai can recommend a starter model, help you download it, and automatically connect it to your first Bees. Download a model. Give it a job. Start. Your models. Your memory. Your device.

What’s New • Smoother, more polished interface We’ve refined several parts of the UI to make morai feel cleaner and more responsive. • Lower power usage Local model performance has been optimized to reduce power consumption, helping your device run more efficiently during longer conversations. • Faster model loading Models now load more efficiently, so you can start chatting sooner with less waiting. • Improved memory performance Memory components have been optimized for faster loading and smoother access to your saved context. • Overall performance improvements We’ve made additional optimizations throughout morai for a faster, lighter, and more reliable experience.

morai can now talk with you — completely on-device. Built-in Voice — Voice is now included with Morai, so there’s no need to download it separately. Real-Time Voice Conversations — Talk naturally with your local AI and hear responses in real time, all processed on your device. Refined UI — Improved parts of the interface for a cleaner, smoother experience. Bug Fixes & Improvements — Fixed several issues and improved overall stability.

• Refined the interface to make everyday use feel cleaner and more intuitive. • Fixed Advanced Voice so you can use voice features reliably again. • Improved overall stability and performance.

• Quick Capture From Home Type directly into the home input. morai recognizes whether you are writing a note or asking a question — saving thoughts to your private journal or opening a local AI conversation when you need an answer.

Morai is no longer just a chatbot, but your local AI workspace. Introducing Memory Map, giving you a clear overview of your AI's memory. It can remember you, organize your projects, read your files, and assist with image processing. A new memory extension system has been added, allowing you to generate and export structured memory via local models for use on any platform.

morai now launches significantly faster, with smoother performance and improved overall stability. This update also includes interface refinements and under-the-hood optimizations to make the experience feel calmer and more reliable.

morai 0.0.8 A major step toward a calmer, more reliable private AI experience. NEW • Detailed Heat & Performance Page Understand how local AI affects your device with expanded thermal monitoring, performance insights, and stability-focused controls. • Smarter File Understanding Models can now work more naturally with local files, helping you chat with documents and project content more effectively. • Automatic Context Compression morai can now intelligently compress long conversations to help reduce memory usage, improve responsiveness, and support longer AI sessions on-device. • New Power Saving Modes Added battery-aware performance options designed for longer AI sessions and reduced device strain. • Refined Interface Updated the UI with a cleaner, calmer, and more focused visual experience across the app. • Smarter Model Recommendations morai now helps recommend models based on your iPhone, available memory, and real-world usability — making local AI easier to understand and manage. • Improved Model Management Better organization, clearer model information, and a more reliable local model experience. • New morai Identity microAI is now morai — a private AI designed to stay with you. • Stability & Efficiency Improvements Reduced power consumption, smoother performance, and multiple optimizations to help reduce crashes during long AI sessions. Not every thought belongs in the cloud.

1. Optimized compressed memory system. 2. Optimized AI response latency, significantly reducing the delay for the initial dialogue. 3. GPU acceleration is disabled by default for better compatibility with older devices, with a new toggle reminder included. 4. Adjusted parts of the UI. 5. Added CPU performance settings to model cards (low-performance mode reduces battery consumption). 6. Added model context length settings. 7. Adjusted chat window UI. 8. Added a Markdown container for responses to standardize AI content delivery.

1. Added project file settings; users can now have their own individual projects and customize Prompts. 2. Adjusted parts of the UI; certain buttons now use native fluid glass UI. 3. Added 2 subscription options; previous "Ad-Free" members are now upgraded to Lifetime Members, enjoying all benefits. 4. Removed all advertisements. Truly local AI. 5. Added project memory functionality. 6. Added the ability to add files to projects, currently supporting txt and pdf formats. 7. Add neural Voice 8.The model download method has been completely modified, allowing users to customize their search and download process.

1. Added two new Gemma 4 models. 2. Added a favorites dialogue component. 3. Adjusted fonts. 4. Updated the Fluid Glass UI. 5. Improved performance.

1,Fixed an issue where new chat windows repeatedly loaded the model. 2,Added a loading animation during AI model conversations. 3,Updated model loading logic for smoother UX performance. 4,Improved model download system — supports resume after interruption and keeps download records after restart. 5,Added enhanced memory system with semantic search and vector matching for more accurate recall.

1. Customizable widgets and memory function. 2. Added support for 8 additional languages. 3. Added support for AdMob pop-up ads. 4. Added a rating system. 5. Improved the UI. 6. Added a remove ad button and a user agreement. 7. Added user instructions.

1. Expand AI Model Selection: Significantly increase the variety of available AI models. 2. Add AI Model Categorization: Implement categories to organize AI models (e.g., by function, size, developer). 3. Local Model Import: Add a feature on the model card to allow users to import their own local AI models. 4. Per-Model Custom Settings: Allow users to configure custom settings for each individual AI model. 5. Improved Chat History Controls: Enhance the management and control of chat history. 6. Display AI Response Speed: Show the generation speed of the AI's response (e.g., tokens per second). 7. Adopt Modern iOS UI: Implement the latest iOS fluid, glassmorphism-style UI design. 8. Show Token Count: Display the token count used in the AI's responses. 9. Add Model Card Actions: Add "Load" and "Settings" buttons directly to the AI model cards. 10. In-Chat Model Switching: Allow users to switch between different AI models directly from the chat screen, without having to return to the model selection page. 11. Update Download Manager: Modify the download utility to improve download speeds. 12. Add Feature Explanations: On the models page, add a pop-up window or modal to explain features. 13. Add Image Recognition (Vision) Chat: Implement chat functionality that can process and discuss images. 14. New User Onboarding: Add a welcome and guidance screen for new users. 15. New App Icon:Replace the current application icon with a new design. 16. GPU Acceleration Toggle: Add a button or toggle to enable/disable GPU acceleration. 17. Model-Categorization Tags: Add tags to models for better sorting and classification. 18. Miscellaneous Model Adjustments: Other minor adjustments and optimizations for models. Bug Fixes 19. Fix AI Model Download URLs: Correct any broken or outdated download links for AI models. 20. Fix Corrupted Models from Interrupted Downloads: Resolve the issue where models fail to load correctly if their download was interrupted. 21. Fix Model Refresh on Restart: Fix the bug where models do not appear or refresh properly after restarting the application.