ComfyUI Extension: comfyui-llamacpp-plus

Authored by pointave

Created

Updated

2 stars

Run ComfyUI workflows without the setup

No installs, no CUDA version roulette, no GPU sitting idle on your bill. Bring a workflow and run it in the browser.

Access llama.cpp server, plus ease-of-life functionality

Looking for a different extension?

Custom Nodes (6)

README

comfyui-llamacpp-plus

A set of nodes to connect llama.cpp models to ComfyUI. Supports text, image, video, and audio inputs, multi-turn chat, HTML/chart visualization, model unloading, and reasoning mode.

If you want a UI to update and load your llama.cpp server and are on Windows, check out Togglellama to have a taskbar system tray icon that unloads model and has 8 custom flag presets.

Nodes

  • LlamaCPP Connectivity — Set your server URL, model, and keep-alive behavior
  • LlamaCPP Chat — Main inference node. Supports multi-modal input, session memory, and HTML visualization
  • LlamaCPP Options — Optional sampling overrides (temperature, top-k, seed, context length, etc.)
  • LlamaCPP Visualizer HTML — Renders any HTML string to a ComfyUI IMAGE via Playwright
  • OpenAI TTS Connectivity -- Refresh will populate dropdown with model deployed
  • OpenAI TTS Speech -- Outputs audio using either reference from audio input or filepath

Installation

cd ComfyUI/custom_nodes
git clone https://github.com/your-username/comfyui-llamacpp-plus

For HTML visualization:

pip install playwright && playwright install chromium

Notes

  • Set keep_alive=0 to unload the model from VRAM after each run
  • media_mode=none lets you leave inputs wired without sending them

License

MIT

Run ComfyUI workflows without the setup

No installs, no CUDA version roulette, no GPU sitting idle on your bill. Bring a workflow and run it in the browser.

Learn more