Openspace3D 1.96

Smarter Imports, AI, and Speech

We are pleased to announce the release of OpenSpace3D 1.96.0. This update brings powerful new character import and animation workflows, deeper AI integration, enhanced speech capabilities, and various performance improvements across the editor and engine.

Import Models – Full Character Workflow

OpenSpace3D now makes it effortless to import complete character models from avaturn.me or any other source, with or without separated meshes. You can also import and merge character animations from files included in the same character folder, perfect for integrating assets from Mixamo or other libraries. New options let you remove root bone translations when needed, and the editor now includes cinematics groups for characters with multiple meshes sharing the same animation.

AI – Smarter, More Flexible

The AI system now uses the response API by default, with an added API type selector to work with v1/chat/completions compatible services. Tool calls are more robust, and we have introduced new actions and events to manage an agent dynamically.

Speech Recognition / TTS – Better Audio Handling

Speech recognition now features automatic input gain management and supports multiple instances for grammar recognition, allowing you to use it as a wakeup word while keeping other recognition tasks running.

Text-to-speech lets you add custom events in your text, such as [happy 0.5] or [dance], to trigger character state changes or animations. Multiple TTS instances are supported, and visemes have been refined for more accurate lip‑sync timing.

PlugITs

Several PlugITs have been updated with new actions and events, including a new Blendshape PlugIT for facial animations, improved animation transitions, and enhanced logging options with custom prefixes and paths.

Editor – Smoother Workflows

Importing xos files is now more flexible: you can bring them into the current scene instead of creating a new resource group. We have also fixed a UI issue with fonts being destroyed too soon, and local documentation is now set by default for faster offline access.

Platform Updates

The core engine has received general performance optimisations, and Android support is now updated to API 37.