Developer Builds Face-Based Avatar Interface for Local AI Agent
A developer has replaced the standard chat interface for their locally-run LLM agent with an animated avatar called Ghost Vessel, designed to sit persistently on-screen like a video call. Instead of displaying plain text replies, the system splits agent output into spoken dialogue, visual data cards, and emotion cues that drive the avatar's facial expressions. Pre-rendered video clips are used for animation rather than live GPU inference, keeping the hardware free to run the language model itself. The avatar integrates with the agent as a standard connector over a local WebSocket, supporting voice input and output, slash commands, and a mood-persistence system. The underlying engine is open-source under the MIT license, allowing others to build their own avatar interfaces by supplying a folder of named video clips.
This is an AI-generated summary. ShortSingh links to the original source for the complete article.
Discussion (0)
Log in to join the discussion and vote.
Log in