nizima LIVE Adds AI Assistant and Voice-Only Tracking in Major Version 2.7 Update
Live2D Inc. has released version 2.7 of “nizima LIVE,” a tracking app for VTubers. This update introduces six new features, including an “AI Assistant” that integrates with chat AIs like ChatGPT and Gemini, and a “Voice Tracking” function that animates models using only voice input without a camera. The update significantly expands the creative possibilities for VTuber activities.
Key Points
- AI Assistant: Integrates with ChatGPT, Gemini, and local LLMs to enable conversations with Live2D models
- Voice Tracking: Camera-free operation — animates eye blinks, lip sync, and full-body movement using voice input alone
- Text-to-Speech: Choose from 31 VOICEVOX characters with 82 voice variations
- Additional Features: Time parameter function, macOS full-model lip sync support, mouse click shortcuts, and Google login
Converse with Your Live2D Model via the AI Assistant
The headline feature, the “AI Assistant,” lets users connect any Live2D model to chat AIs including ChatGPT, Gemini, and local LLMs for interactive conversations. For text-to-speech, 31 VOICEVOX characters with 82 voice variations are available, including popular characters like Zundamon and Shikoku Metan. The model’s mouth and body move in sync with the speech, and emotions inferred from the conversation content are reflected in facial expressions. Responses can also be output as subtitles to OBS, making it possible to display the AI assistant’s dialogue on stream.

Camera-Free Voice Tracking
Another standout feature is “Voice Tracking.” Without any camera-based tracking, the system naturally controls eye blinks, lip sync, head tilting, and full-body movement using voice input alone. This is particularly useful when users want to animate their model without live footage or when a camera setup isn’t available. It also works great during art streams when the creator is looking down, making camera tracking difficult.

Packed with Additional New Features
Version 2.7 brings a host of other additions as well. The “Time Parameter” function lets model effects change according to real-world time. The macOS version now supports lip sync for all models, bringing it on par with the Windows version. Users can also assign expression and motion shortcuts to mouse clicks, and Google account login is now supported. New effects including “Mirror Ball,” “Comic,” and “Pixel Art” have also been added.

Future Outlook
Live2D Inc. plans to continue updating the AI Assistant feature, proposing its use as a chat companion for casual streams or as a character to liven up broadcasts. nizima LIVE can be downloaded and updated from the official website, and the “nizima LIVE TRACKER” is also available for iOS. Combined with the Live2D model marketplace “nizima,” the ecosystem supporting VTuber activities continues to grow.
Commentary and Prospects
The integration of an AI Assistant into nizima LIVE is a landmark update that signals the future direction of VTuber tools. With growing interest in Live2D technology, including the recent publication of Live2D-related books, the AI chat integration opens up the possibility of “character-to-character banter even in solo streams” — a entirely new streaming format. The Voice Tracking feature is also a long-awaited addition for creators constrained by their camera setup. nizima LIVE, accessible to beginners and professionals alike, continues to lower the barrier to entry for aspiring VTubers.
Source: PR TIMES
