1.8 KiB
1.8 KiB
Text-Driven Prompting Workflow Guide
This workflow streamlines prompt management and eliminates manual weight tweaking by utilizing two custom nodes for layered control.
Node Overview
1. Text Line Prompt (Text-Driven)
Manages a list of prompt assets or concepts line by line.
- Bi-directional Sync: Select a line via the dropdown menu to modify its multiplier or active state. The raw text area updates automatically in real-time.
- Muting Function: Deactivating a line prepends
//to comment it out. These lines are completely skipped during backend execution to output a clean string. - Regex Parsing: Properly handles character LoRA escaped brackets
\( \)and safely bypasses rows containing multiple unescaped brackets.
2. Multi Prompts (Text-Driven)
Provides centralized layer-control for multiple modifiers.
- Dynamic UI Layout: Input fields dynamically hide or show based on the "num_prompts" value (up to 10 slots) to keep the canvas clean.
- Granular Controls: Allows you to adjust individual multipliers and toggle the active state of each prompt fragment with a single click.
Workflow Pipeline
The nodes are connected in a relay structure to compile the final prompt:
- Base Generation: [Text Line Prompt] manages base characters, core concepts, or specific styles.
- Modifier Stacking: The selected base prompt is fed into [Multi Prompts] via the hidden
optional_text_ininput pin. - Final Output: [Multi Prompts] appends aesthetic modifiers (lighting, attire, quality tokens), then sends the fully combined string directly to the SDXL CLIP Text Encode.
Quick Start
- Choose a row in Text Line Prompt using the dropdown menu.
- Tweak individual weights or toggle active states in Multi Prompts.
- Click Queue Prompt to execute.