Open-LLM-VTuber

Why we included this project

Open-LLM-VTuber gives any LLM a face and a voice: speech recognition, text-to-speech, and a Live2D avatar that reacts as you talk. You can interrupt mid-sentence, it responds to touch, and a desktop pet mode keeps the character sitting on your screen when you want company around. Because the language model and the voice stack can both run locally, the whole setup works offline, which matters if you don't want audio leaving your machine. Hobbyists building an AI waifu, virtual pet, or streamer-style avatar get a long list of pluggable pieces (Ollama, OpenAI-compatible endpoints, Whisper variants, several TTS engines) and can swap them freely, while the included web and desktop clients mean you don't have to assemble the pipeline yourself.