ComfyUI Extension: ComfyUI Grok Prompts

Authored by babydjac

Created

Updated

1 stars

Custom ComfyUI nodes for generating prompts using Grok AI, enhancing prompt creation for text-to-image workflows.

Custom Nodes (0)

    README

    🧠 ComfyUI Grok Prompts

    AI-Powered Prompt Engineering Nodes for ComfyUI using xAI's Grok API.
    Fine-tune your prompt craft with LLM intelligence, split, recombine, and visualize prompt transformations, all inside ComfyUI.

    🔧 What is this?

    comfyui-grok-prompts is a modular node suite for ComfyUI that uses xAI’s Grok LLM API to transform user input into optimized prompts for AI models like PonyXL and Flux. These nodes intelligently:

    • ✂️ Split prompts into subject/scene (text_l) and style/mood (text_g)
    • 🔀 Combine and rewrite them into perfectly scored Danbooru-tagged prompts
    • 🎥 Output video prompts using Wan’s Advanced Prompt Formula
    • 💬 Display a UI explanation of how the prompt was refined

    🌈 Live Visual Demos

    | Model | Image | Video Preview | |-----------|-------|----------------| | PonyXL | Pony | <video src="https://raw.githubusercontent.com/babydjac/comfyui-grok-prompts/main/media/PonyVideo.mp4" autoplay loop muted playsinline width="320"></video> | | Flux | Flux | <video src="https://raw.githubusercontent.com/babydjac/comfyui-grok-prompts/main/media/FluxVideo.mp4" autoplay loop muted playsinline width="320"></video> |

    🧩 Node Packs Included

    🔹 PonyXL Grok Prompter

    • 🔤 Takes your prompt and transforms it into:

      • ponyxl_prompt: score_9, score_8_up, rating_mature, 1girl, brunette_hair, long_hair, beautiful_face, detailed_eyes, blue_eyes, full_lips, blush, gigantic_breasts, cleavage, casual_clothing, shopping_mall, indoor, waist_up
      • wan_prompt: a video of a gorgeous brunette walking through a busy mall, focusing on her face and upper body

    • 💬 Also displays Grok's explanation:
      “Combined and cleaned Danbooru tags for clarity, emphasis, and aesthetic scoring. Generated concise video scene prompt from subject + motion.”

    🔸 Flux Grok Prompter

    • 🔤 Optimizes prompts for abstract or high-concept visuals (Flux model style):

      • flux_prompt: a close-up view of a stunning brunette woman walking through a bustling mall, surrounded by vibrant shadows and colors
      • wan_prompt: a video of a brunette woman walking confidently through a mall, with camera focused on her face

    • 💬 Grok's explanation:
      “Emphasized human detail and cinematic energy; prompt pruned and rewritten for maximum Flux fluency.”

    📂 Media Folder Contents

    All media live in /media:

    | File | Description | |------|-------------| | PonyExample.png | PonyXL still image output | | FluxExample.png | Flux still image output | | PonyVideo.mp4 | PonyXL video animation | | FluxVideo.mp4 | Flux video animation |

    🎨 UI/Visual Assets

    Supporting visuals:

    • Asset 1
    • Asset 2

    🔐 xAI API Required

    To use these nodes, you’ll need an API key from x.ai. Input your key into the node’s api_key field.

    🚀 Use Case Highlights

    • Perfect for NSFW image prompting using PonyXL
    • Abstract stylized generation with Flux
    • Natural-language video scene building for Wan
    • Ideal for artists, animators, and prompt engineers wanting more automation and precision

    🫡 Credits

    Brought to you by @babydjac
    Powered by: ComfyUI + xAI’s Grok

    Prompt smarter. Generate harder.