Open-LLM-VTuber

Open-LLM-VTuber is an AI companion for people who want to talk with a customizable animated character. Local language and speech models let it work fully offline, with conversations staying on your device.

• Character conversations: Talk by voice or text with a Live2D character, interrupt its speech, and let it speak proactively; its appearance and persona are customizable.
• Visual interaction: Give the companion camera or screen input, interact through clicks and drags, or use group chat; MCP support and browser use extend what the AI can do.
• Desktop pet: Keep a transparent character on your desktop with movable placement, an always-on-top mode and mouse click-through.
• Conversation continuity: Revisit saved chats and use memory integrations such as Letta or EVI, with support for custom memory systems.

Web and desktop clients run on Windows, macOS and Linux. Local language backends include Ollama, LM Studio, vLLM and GGUF models. Speech options include Whisper.cpp, Faster-Whisper, Coqui-TTS, GPTSoVITS and CosyVoice; speech translation lets the character reply aloud in a different language.

CPU operation and NVIDIA or non-NVIDIA GPUs are supported, with macOS GPU acceleration available for some components. Optional services such as OpenAI, Claude and Gemini process the requests sent to them outside your device.