Knowledge 2.0. Inworld’s new knowledge system is now live, offering improved knowledge retrieval, auto-generation of knowledge from documents, and the use of tags to organize content, with all functions consolidated into a single UI.
Voice Studio for Chinese, Japanese, and Korean. Voice Studio now supports Chinese, Japanese, and Korean, allowing these workspaces to work with speakers of any of Inworld’s supported languages.
Multi-Agent Chat. Studio now supports chat between multiple characters, allowing you to test how characters interact with each other both with and without player input.
Character and Workspace Search. Inworld studio now supports searching across workspaces, characters, and knowledge, resulting in much quicker navigation.
Unreal/Unity: Inaccurate Response Time Reporting. The Reported Response time for Unreal and Unity now accurately reflects the Player Experienced response time.
NodeJS: First Character Response Not Printing. Users will now see the first line of a character’s response.
Unreal: Dropped Info on Session Pause. Pausing a session after sending a trigger no longer causes info loss.
Multilingual AI Models. It is now possible to create Mandarin (Simplified Chinese), Japanese, and Korean workspaces to build characters that support these languages.
Multilingual Voices. A new set of Mandarin (Simplified Chinese), Japanese, and Korean voices are now available, which are accessible via both the language-specific workspace and Inworld’s standalone voice API.
Multilingual Speech-to-Text. Mandarin (Simplified Chinese), Japanese, and Korean are now supported in Speech-to-Text within the associated language workspaces, allowing native speakers to communicate with characters via mic in those languages.
Intent Recognition System. Improvements have been made to intent detection within the Goals feature to reduce false positives and better differentiate between similar intents, resulting in noticeable behavioral improvements.
Voice Studio. Without needing to create a character, this feature allows you to seamlessly listen to Inworld’s full library of diverse voices, adjust pitch or talking speed, and download samples, all with the click of a button.
YAML Goals Visualizer. A new set of shortcuts, copying and pasting triggers, and node visualizations are now available when enabling or disabling Goals Mutations, allowing you to work faster and with more clarity when creating Goals and Actions for your characters.
Studio Onboarding Tour. Experience a guided tour of Inworld Studio, with step-by-step guidance for key actions, such as creating and editing a Character. Tours can be triggered by going to: Profile Settings → Manage Account → Launch Tour.
Updated Integration Tab UI. This updated tab offers cleaner access and management of the Inworld Studio API and Runtime API keys from the Integrations tab in Inworld Studio.
Legacy Voice Deprecation. Thanks to the improvements with Inworld’s new voices, the legacy voice library has been deprecated and characters using the older voices have been automatically migrated to either their new updated voice equivalents or a default voice at no cost. Existing characters can be switched to use any of the new voices as needed.
Voice Pitch Slider. Due to popular demand, we have introduced the ability to change the pitch of individual voices for Inworld’s new voices, which brings back a functionality that existed with the older legacy voices.
Enhanced Code Editor. Experience a full screen YAML editor when configuring your game mechanics, with side-by-side sample code and documentation.
Studio Access Token Deprecation. Starting from Unity SDK 3.5 and Unreal SDK 1.7, the Studio Access Token is being deprecated and replaced with Studio API Base64, for the purposes of logging into the Inworld Panel in Unity and Unreal SDKs. This change is being made to ensure user account security.
Chat Bubbles No Longer Remain: Due to a race condition, chat bubbles were displaying when canceling a response.
Unity: Audio Exception Fixed: An audio capture exception no longer occurs when the mic is set to 44.1khz.
Unity: Blurry Shadows and Textures. Sample scene RPM Characters no longer display blurry shadows and textures. Previously, ReadyPlayerMe characters were displaying low-res textures and shadows, causing them to appear blurry.
Unity & Unreal: Lip-sync and 2.0 Voices. Lip sync no longer continues through to the end of utterances when using Inworld's new 2.0 voices. Before, lip-sync could cut off several seconds before the audio was finished.
Unreal: Scaling Inworld Images. Images are now correctly scaled in the Unreal editor plugin. Before, logos and character images would not scale with the size of the editor window.
Node-js: Missing onMessage Callback. Added missed onMessage callback to the change scene method.
Web: Push-to-Talk. An error involving using both push-to-talk and open mic in the same session has been fixed. Previously, if both push-to-talk and open mic were used in a single session, an error would occur in the console.
Godot SDK. We are thrilled to announce the release of our open-source Godot SDK. It's now available for developers and enthusiasts to explore and contribute.
Documentation. Documentation can be found both in the repository's README and on our docs site, providing guidance for using the SDK.