Release v6.0

Updates:

• Fully rewritten UI
• Added auto-scaling
• Mobile Optimization
• Scaling optimization
• Full dark mode support
• New sending message logic
• Implemented prompt templates
• Himitsu implemented in the as prompt template manager (for now)
• Workaround single-page application
• Full PWA support
• The sidebar was returned
• Sidebar mobile optimization
• Select elements now are shown as a scrollable lists
• Message history elements are now showing correctly in full-screen
• Server loading delay was decreased

New

• Parameters on client configuration
• Server side in file configuration
• Image generation
• Image recognition
• Follow-up copilot for prompt templates
• Icons instead of text buttons
• Yuna logo
• Call UI
• Loading animation
• Auto fix option for server deploy
• Helpful automation configuration shell script

Fixed:

• Sounds error
• Wrong windows of popup elements
• Full-page freeze

Not working:

• Speech recognition
• Speech synthesis

Not implemented

• Web search
• Long-term memory embeddings

Full Changelog: v5.2...v6.0

Release v5.2

Full Changelog: v5.1...v5.2

What's new:

• New mobile-optimized UI
• Rewritten UI for single-page application
• Implemented Yuna's vision, one more step towards AGI
• Image captioning and image generation were moved to the separate vision module
• Project file structure logic was refactored
• Better generation with native Python libraries
• First stable workaround for custom configs
• All C++ dependencies were removed
• New sounds
• Improved server request logic
• Implemented full video call support
• Images now integrated into message bubbles
• Easier installation
• Memory efficient generation

Still not working:

• TTS response with audio
• PWA is not working offline but working in general
• Speech recognition causes error: "can't recognize the text."
• Speech recognition not working on mobile
• Config not loading from external URLs

Release 5.1

Full Changelog: v5.0.1...v5.1

What's new:

• Implemented video call support with image recognition using diffusion description
• Improved UI
• Updated UI to Kawai v5.1
• Fixed menus
• Improved mobile adaptivity
• Stabilized server works
• Full client solution
• Better dark mode
• Fixed overlaying CSS
• Project partial factoring
• Small directory renamings
• Uploaded datasets in the first beta
• Moved libraries in the "lib/" directory
• Fixed request bug
• Fixed overlaying topbar
• Improved voice recognition

Still not working:

• TTS bug when getting a response
• Image description not implemented into dialog
• PWA is not working offline

Release 5.0.1

Improved:

• Message sending
• Request waiting window
• Popups and configuration
• UI
• Documentation
• Updated llamacpp support for higher speed

Still not working

• Video calls
• Fine-tuning mechanism
• Correct multimodel redirection
• Long-term memory is not implemented

Release 5.0

• Improved general stability
• Rewritten UI
• Refactored logic from scratch
• Client and Server split
• Moved to clear HTML implementation
• Dymanic server config
• Faster loading
• Less VRAM usage
• Faster model loading

Release v4.0

• Improved stability
• Full refactoring of main function
• Video calls will be in v4.1

Release v3.3

• History management
• Call

Release v3.1

• Better generation
• X2 speed
• Voice call

Release v3.0

New features of of Yuna:

1. Support for Video and Audio Calls:
   Yuna AI offers a native web UI for seamless video and audio calls, enhancing your communication experience.

2. Image Processing and Creation:
   Yuna can process images and create drawings, adding a creative touch to your interactions.

3. Emotion Understanding:
   Yuna has a full range of emotions, allowing her to understand and display emotions realistically.

4. Updated Model:
   Yuna AI utilizes a powerful model with 7 billion parameters, ensuring her responses are insightful and engaging.

5. Hardware Acceleration:
   Yuna supports CUDA GPU and CPU MacOS MPS acceleration for smooth performance.

6. Native Telegram API:
   Communicate with Yuna through the native API, enabling calls and text interactions via Telegram.

7. Web App Support (PWA):
   Yuna AI can be accessed as a Progressive Web App (PWA) for convenient use across different devices.

8. Kawai Framework:
   The CSS and JS of Yuna's interface have been rewritten using the Kawai Framework for a visually appealing design.

9. Platform Compatibility:
   Yuna AI is adapted to run on Linux, MacOS, and Windows, ensuring cross-platform support.

Release v3.0 ~alpha 2

Optimized model