ComfyUI-Muse
ComfyUI-Muse embeds a persistent, multi-session local LLM chat panel directly inside a ComfyUI node, so a user can brainstorm image or video prompts against Ollama, LM Studio or a directly loaded GGUF model without leaving their generation workflow.
The panel keeps several independent chat sessions, each with its own history, system prompt, model and settings, saved to disk and restored across restarts, and streams replies token by token with a stop button to cancel generation. A direct model loader can run GGUF models straight from disk without LM Studio or Ollama running at all, and the panel coordinates VRAM between chat and image generation so a user does not have to remember to unload one before running the other.
Per-chat reference files can be attached to steer every message in a session toward a house style, and vision-capable models can be given attached images or, for capable models, sampled video frames to discuss.
The only hard dependency for the core chat panel is aiohttp, which already ships with ComfyUI, and the project is released under the MIT license.