ComfyUI-WanPromptBuilder

Three ComfyUI nodes for generating structured WAN 2.2 video prompts , built specifically for chained multi-segment video pipelines.

---

Nodes

WAN Prompt Builder (Groq)

Single prompt for a single video segment. You give it a subject, action, camera move, lighting, style and intensity , it returns one clean WAN-optimized prompt via Groq/Llama.

WAN Prompt Builder Trio

For long-form video generation across 3 chained segments (~7s each). Generates 3 coherent prompts from a single set of inputs, with automatic camera progression (wide > medium > close). Same scene, continuous action, no jarring cuts.

WAN Prompt Builder Vision

Same as Trio but takes an image as input as visual reference. Automatically describes the image and builds the 3 prompts around it. Useful when you're animating an existing still or concept art.

---

Features

• WAN 2.2-specific prompt rules baked in (one action per segment, no action chaining, correct sentence structure)
• Action intensity slider (1–5) controls energy level across all prompts consistently
• Camera move and lighting dropdowns with WAN-friendly presets
• Fallback prompts if Groq API times out , never breaks your workflow
• Saves API key to config.json so you don't re-enter it every time

---

Requirements

• ComfyUI
• A Groq API key , free tier works fine, takes 2 minutes to set up at console.groq.com

---

Installation

cd ComfyUI/custom_nodes
git clone https://github.com/YOUR_USERNAME/ComfyUI-WanPromptBuilder

Restart ComfyUI. The nodes appear under WAN > prompting.

---

Why Groq?

Fast, free tier is generous enough for personal use, and the models (llama-3.3-70b-versatile for text, llama-4-scout-17b for vision) produce solid cinematic prompt language out of the box.

---

Support

If this saved you time → Buy me a coffee