Prompt Assistant

Prompt Assistant

MooshieUI ships a local LLM (a llama.cpp GGUF server) that helps you write prompts. It runs on your own machine, so prompts are not sent to any external service.

What it does

• Enhance - rewrites and improves your existing prompt.
• Compose - generates a prompt from a plain-language description. You choose a length (short, medium, or detailed) and whether to include artist tags.

How it runs

• The assistant starts a local llama.cpp server on demand.
• It is hardware-aware: it offloads layers to the GPU when there is room and falls back to CPU if ComfyUI is already using your VRAM.
• An idle watchdog stops the server automatically when it has not been used for a while, freeing memory.

Hosted / Docker notes

• In GPU Docker builds you can point the assistant at a specific llama.cpp binary directory with the MOOSHIEUI_LLAMA_BIN_DIR environment variable.
• On multi-GPU hosts you may want to pin the llama server to a specific device (for example CUDA_VISIBLE_DEVICES) so it does not compete with ComfyUI.
• The model used by the assistant is configured by the deployment. If no assistant model is configured, a fallback tag-generation model may load instead.

See also

• Prompting Guide for wildcards, presets, scheduling, and the rest of the prompt system.