GPT-VUP is an open-source project published on GitHub that turns ChatGPT into a virtual streamer (VTuber) able to host live broadcasts on Bilibili and Douyin. The project has gathered around 1,300 GitHub stars.
GPT-VUP connects the pieces an AI livestreamer needs into one pipeline: it listens to the live room, feeds danmaku (bullet comments) and prompts to a GPT model, and speaks the replies aloud with text-to-speech while an animated avatar performs on screen. Speech recognition lets the virtual host react to voice as well as text chat, and embedding-based memory helps it keep track of conversation context and recurring viewers. The project became a reference implementation during the first wave of AI VTubers in 2023, when fully automated streamers started appearing on Chinese platforms, and its code shows how to wire LLM chat, TTS, live-comment ingestion and avatar presentation together end to end. The repository was last updated in 2023; running it requires your own OpenAI API key and the usual streaming setup for your platform.
GPT-VUP is free and open source. You only pay for your own OpenAI API usage and any TTS services you plug in.
Streamers, VTuber hobbyists and developers curious about building an autonomous AI livestreamer for Chinese platforms.