ChatdollKit

Why we included this project

ChatdollKit is a Unity SDK that gives a 3D avatar the ability to hold spoken conversations, which is useful if you're building a character with a visible body and voice rather than a plain text interface. It handles the entire loop: speech-to-text, a language-model brain that supports ChatGPT, Claude, Gemini, or Dify (with function calling and multimodal input), and text-to-speech from OpenAI, Azure, Google, or VOICEVOX. Beyond the pipeline, it also manages lip-sync, facial expressions, blinking, wakeword detection, and barge-in so users can interrupt the avatar mid-sentence. Because it runs on Unity, the same model can be deployed to Windows, iOS, Android, VR, AR, and WebGL. That combination makes it a strong choice for virtual companions, AITubers, or game NPCs where the character's expressiveness is part of the product.