Open-LLM-VTuber
A local AI companion for Windows, macOS and Linux with voice chat, Live2D avatars and camera input. Runs offline with local models or connects to cloud APIs.
Open-LLM-VTuber is a local AI companion for people who want a character they can talk to, with a Live2D avatar that responds through speech and expressions. It runs on Windows, macOS and Linux through web and desktop clients. With local models for speech and language processing, it works fully offline and keeps conversations on your device. Cloud APIs are optional alternatives that send the corresponding processing to external services.
You can choose the character's appearance and persona, then interact through voice or text. You can interrupt its speech, and it can speak without waiting for a prompt. Camera and screen input let it respond to what it sees; touch interactions and group chat add other ways to engage with the character. It also supports MCP and lets the AI use its own browser.
The desktop client includes a transparent pet mode with a movable avatar, an always-on-top option and mouse click-through. Conversation history lets you return to earlier chats. Optional agents can connect to services such as Letta and HumeAI EVI; these require separate configuration.
Backend choices include Ollama, LM Studio, vLLM and GGUF models, alongside OpenAI-compatible APIs and services such as Claude and Gemini. Speech recognition options include Whisper.cpp, Faster-Whisper and sherpa-onnx; speech synthesis includes Coqui-TTS, GPTSoVITS and CosyVoice. Speech translation lets the character answer in a different spoken language. CPU operation and NVIDIA or non-NVIDIA GPUs are supported, and some components can use GPU acceleration on macOS.
The code is MIT-licensed. Bundled Live2D sample models have separate license terms. Version 2 is still in planning; the available version 1 continues to receive bug fixes.
