slahiriandClaude Opus 4.5
0dfd6fd294
New node that modifies prompts based on Z-Image structure: - Takes input prompt and LLM model - 6 section instruction inputs: subject, clothing, pose, environment, lighting, camera - Two processing modes: Single Pass (faster) and Section by Section (precise) - Temperature control for LLM creativity - Optional Instagram caption generation (3 styles) - Text-only LLM calls (no image required) Phase 1 implementation using LLM API. Phase 2 will add SID_LLM_Local_Text support. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
ComfyUI-AI-Photography-Toolkit
AI-powered prompt generator for ComfyUI. Analyzes images and generates detailed prompts optimized for Z-Image Turbo and other image generation models.
Installation
cd ComfyUI/custom_nodes
git clone https://github.com/slahiri/ComfyUI-AI-Photography-Toolkit.git
Restart ComfyUI. All dependencies install automatically.
Nodes
| Node | Screenshot | Description |
|---|---|---|
| SID_LLM_Local |  |
Run vision models locally (Qwen3-VL, Florence-2, etc.) - no API key needed |
| SID_LLM_API |  |
Use cloud providers (Claude, GPT-4o, Gemini, Grok, etc.) |
| SID_ZImagePromptGenerator |  |
Generate prompts from images |
Quick Start
Option 1: Cloud API (Best Quality)
- Add SID_LLM_API node
- Select provider (e.g.,
Anthropic) - Enter your API key
- Add SID_ZImagePromptGenerator node
- Connect:
LLM node → Prompt Generator ← Image - Run!
Option 2: Local (Free, No API Key)
- Add SID_LLM_Local node
- Select model (e.g.,
Qwen3-VL-8B-Instruct) - Add SID_ZImagePromptGenerator node
- Connect:
LLM node → Prompt Generator ← Image - Run!
Workflow
Image ─────────────────┐
├─→ SID_ZImagePromptGenerator ─→ prompt ─→ CLIP Text Encode
SID_LLM_API ───────────┘
(or SID_LLM_Local)
Supported Providers
Cloud (require API key): Anthropic Claude, OpenAI GPT-4o, Google Gemini, xAI Grok, Mistral, Together AI, Fireworks, Groq, OpenRouter
Local (free): Ollama, LM Studio, or use SID_LLM_Local with built-in models
Sample Workflow
Download: sample_workflow.json
Tips
- Best quality: Use Claude or GPT-4o with "Enable Reasoning" turned on
- Best local: Use Qwen3-VL-8B-Instruct with reasoning off
- Fastest: Use "Quick" analysis mode
Version: 4.2.0 | Author: Siddhartha Lahiri
For advanced configuration, provider details, and migration guide, see Technical_Details.md.