Voxta 149 Beta: 3D Avatars, MCP, Collections & Massive UI Upgrades!

Hey everyone,

The new Voxta 149 Beta is now available. This update focuses on adding significant new features, like support for 3D VRM avatars and AI-assisted document editing, alongside many important UI and backend improvements.

✨ New Features & Major Updates

• 3D VRM Avatar Support (Experimental): You can now use 3D VRM avatars for your characters. Add a .vrm file to the character's assets, along with an Idle.fbx (currently Mixamo only) for animations. Your character will appear in the UI and can be animated.

• AI-Assisted Document Editing (Experimental): We've added a new way to work with text. You can drop content into a shared canvas that your AI character can see and read, allowing for collaborative editing and analysis.

• Expanded MCP Server Support: Voxta can now connect to more external tools through our Model Context Protocol (MCP), including support for images and arguments. This allows characters to trigger a wider variety of actions.

• UI Sliders for LLM Parameters: We've added UI sliders for settings like temperature, so you can adjust LLM parameters more easily without having to manually input numbers.

• Collections System (Experimental): You can now group characters, scenarios, and other assets into a single "collection." This is an early feature that we plan to build on for easier content sharing in the future.

⚙️ UI and Workflow Improvements

• New Gallery & Folder System: We've added new gallery and table views for Browse scenarios and memory books, now with the ability to organize items into folders.

• Easier Comfy UI Integration: We've worked on simplifying the use of Comfy UI workflows. For many setups, you can now just drop the exported .json file into the Voxta Comfy UI folder.

• Improved Assistant View: The assistant view has been updated with better text animations and a cleaner layout.

• Scripting Updates: Scenarios and characters can now generate images directly using the new chat.generateImage() command.

• General UI Polish: We've improved dropdown menus, hover effects, and the character grid layout. We also added video avatar support in the chat tab and better organization for LLM settings.

🛠️ Key Fixes

• Fixed an issue where cloning a scenario would not copy its assets.

• Addressed a memory leak related to audio references that could cause audio to stop working over time.

• Corrected a bug where chat variables weren't saving correctly after being modified by a trigger.

• Fixed image caching issues that could sometimes result in the wrong image being displayed.

• Resolved crashes related to ComfyUI image reuse and long seeds in KoboldAI.

• Updated key dependencies like Torch, ExLlamaV2, and Coqui.
  
  Links:
  How to install Voxta server app: https://youtu.be/1I9VkJ8tTlo
  How to update Voxta server app: https://youtu.be/5aa7sducwoc
  

We're looking forward to hearing what you think about these changes. Your feedback is very helpful for us as we continue development.