Voxta

Voxta 150 Beta: New Installer, Configurations, Kitten TTS & Major UI Overhaul

Added 2025-08-28 00:51:19 +0000 UTC

Hello everyone,

Voxta 1.0.0-beta.150 is now available. This update introduces significant improvements to core functionality and the user interface.

A major highlight is the new installer and updater, which simplifies the process of installing and maintaining the application. You no longer need to manually copy files to preserve your settings between updates.

This release also adds a Configurations system. This feature allows you to create and save different sets of services and settings, making it easier to switch between setups for roleplay, assistant use, or other tasks.

We have also added support for new voice options, including ElevenLabs v3 models and the new Kitten TTS, a very lightweight 25MB text-to-speech model.

The full list of changes is below.

✨ New Features & Major Updates

New Installer & Updater: Voxta now includes a full installer and updater for a more straightforward setup and update process.
Configurations System: Create, save, and assign specific service configurations to characters or scenarios, allowing you to easily switch between different setups.
Kitten TTS Support: Integrated a new 25MB text-to-speech model, designed for efficiency and speed.
ElevenLabs v3 Support: Added support for the latest TTS models from ElevenLabs, including their capability to generate non-speech sounds.
Chat-Specific Long-Term Memory: You can now enable a dedicated long-term memory for an individual chat, which will override the character's default memory.
New "Visual Story" Scenario: A new built-in scenario that generates an image after each reply, intended for creating visual narratives (requires an image generation service).
Advanced MCP Actions: The Model Context Protocol (MCP) now supports action grouping and custom layers, allowing for more organized and capable actions with external tools.

⚙️ UI and Workflow Improvements

Overhauled Events & Actions Editor: The editor has been redesigned with a new navigation panel that displays one event at a time. This improves performance and makes managing complex scenarios easier.
Interactive 3D Previews: You can now preview 3D characters and animations (VRM and FBX files) directly in the Character Assets tab, which includes automatic model centering.
Updated Modules & Services Menu: The Services, Modules, and Add Modules pages have been updated with module icons and logos for better usability.
Advanced Prompt Editor for Image Gen: The prompt editor, which supports weight adjustments with Ctrl+Up/Down, is now available in the Image Generation playground.
Assistant View Polish: The Assistant view has been improved with an expandable input box, a smoother typewriter effect, and an updated UI for code blocks and message editing. You can also CTRL+Click to delete a message without confirmation.
In-Editor Voice Testing: Voices can now be tested directly from the character's voice selection menu.
Drag-and-Drop Import: Import .voxpkg files by dragging them from any folder on your computer onto the Voxta window.
Module Management: You can now stop unused modules via the right-click menu in the "Running Modules" panel. Modules without a definition are also shown to allow for easier cleanup.

🛠️ Key Fixes & Technical Enhancements

Dependency Updates: Updated Torch to 2.8, LlamaSharp to 0.25.0, and ExLlamaV3 to 0.0.6.
Model Compatibility: Improved support for models that generate "thinking" tokens multiple times in a single response and added support for vLLM thinking tokens in OpenAI Compatible services.
Audio Stability: Addressed bugs related to audio playback, including "Failed to play audio" errors and issues when playing voice samples in rapid succession.
UI Fixes: Corrected a number of UI bugs, such as system prompt field overflows, unwanted spacing in roleplay text, and the main drop zone being triggered incorrectly when dragging items within the application.
Desktop App: The console now includes buttons to clear the view and export logs.

Links: