comfy-intelliprompt

intelliPrompt for ComfyUI

An intelligent prompt optimizer for ComfyUI that automatically fixes typos, balances parentheses, enhances prompts with quality tags, and optionally uses AI for advanced prompt optimization.

Features

✅ Typo Correction - Fixes common English typos automatically
✅ Parentheses Balancing - Fixes unbalanced (), [], {}
✅ Punctuation Fix - Ensures proper ending punctuation
✅ Quality Enhancement - Adds quality tags (masterpiece, best quality, etc.)
✅ AI-Powered Mode - Uses Pollinations AI for advanced prompt enhancement when optional API dependencies are available
✅ Offline Mode - Works completely offline with local processing
✅ Graceful Fallback - Missing API/runtime deps no longer block node registration; local mode still works
✅ Non-English Support - Translates prompts to English via AI
✅ Lazy Torch Use - The latent preset node registers even if torch is unavailable and only errors when executed

Installation

Fresh install with comfy-sage

For a new comfy-sage install, the supported base/local flow is:

1. Bootstrap comfy-sage once with ./launch-comfy.sh
2. Clone comfy-intelliPrompt into ComfyUI/custom_nodes
3. Relaunch ./launch-comfy.sh
4. Use the node immediately in local mode without any manual pip install step

# from the unpacked comfy-sage directory
./launch-comfy.sh

# after the first launch completes
git clone https://github.com/galpt/comfy-intelliPrompt.git ComfyUI/custom_nodes/comfy-intelliPrompt

# relaunch so ComfyUI registers the node
./launch-comfy.sh

API mode is optional. If its extra dependency is not present or the API is unavailable, intelliPrompt still registers and falls back to local processing.

Standard ComfyUI install

cd /path/to/ComfyUI/custom_nodes
git clone https://github.com/galpt/comfy-intelliPrompt.git

Restart ComfyUI after cloning.

Option 2: Copy folder

1. Download or copy the comfy-intelliPrompt folder
2. Paste it into your ComfyUI's custom_nodes/ directory
3. Restart ComfyUI

Requirements

• ComfyUI
• Python 3.10+
• No extra Python packages are required for base/local prompt cleanup
• Optional API enrichment only applies when its runtime dependency is available
• torch is only required when executing IntelliPromptResolutionPresetLatent
• comfy.model_management is only used when available inside a ComfyUI runtime

Usage

1. After installation, find the "✨ intelliPrompt" node in the node list (under "utils" category)
2. Add it to your workflow between your prompt input and CLIPTextEncode
3. Configure:
   • prompt: Your input prompt
   • optimizer_seed: Random seed (for AI mode consistency; protected from ComfyUI's special seed handling)
   • use_api: Enable/disable AI-powered processing
   • preserve_negative_terms: Conservative cleanup path for negative prompts
   • quality_tags: Additional tags to add (optional)

Node Inputs

| Input | Type | Default | Description | |-------|------|---------|-------------| | prompt | STRING | "A beautiful sunset..." | Your input prompt (positive or negative) | | optimizer_seed | INT | 42 | Random seed for AI API calls | | use_api | BOOLEAN | True | Use Pollinations AI (True) or local only (False) | | preserve_negative_terms | BOOLEAN | False | Conservative local cleanup for negative prompts; disables enrichment/API | | quality_tags | STRING | "masterpiece, best quality..." | Additional quality tags to add (optional) |

Node Outputs

| Output | Type | Description | |--------|------|-------------| | optimized_prompt | STRING | The processed/optimized prompt |

Workflow Example

[Prompt Input] --> [intelliPrompt] --> [CLIPTextEncode] --> [KSampler]
                    (node 21)          (node 6)             (node 3)

How It Works

Local Processing (always available)

When use_api=False, optional API support is unavailable, or the API fails, intelliPrompt performs:

1. Typo Correction: Fixes 20+ common English typos
2. Parenthesis Balancing: Ensures ( ), [ ], { } are balanced
3. Punctuation Fix: Ensures prompt ends with ., !, or ?
4. Space Cleanup: Removes multiple spaces, fixes spacing after punctuation
5. Quality Enhancement: Adds quality tags if none present

AI Processing (optional)

When use_api=True and optional API support is available, intelliPrompt sends the prompt to Pollinations AI which:

• Translates non-English to English
• Adds rich sensory details (lighting, mood, colors, atmosphere)
• Expands style references and artist mentions
• Avoids repetitive vocabulary
• Optimizes for text-to-image models

If the API is unavailable or fails, it automatically falls back to local processing.

Runtime compatibility notes

• __init__.py avoids top-level imports of optional/runtime dependencies so missing extras do not break node registration.
• intelliPrompt always keeps a local fallback path available.
• IntelliPromptResolutionPresetLatent stays registered even in reduced environments; it raises a clear runtime error only if executed without torch.

Why intelliPrompt?

Unlike other prompt optimizers, intelliPrompt:

1. Keeps prompt cleanup available when API mode fails - Falls back to local processing when API optimization is unavailable
2. Works offline - Local mode doesn't need internet
3. Fast processing - Local mode is instant
4. Easy to use - Simple node interface
5. Open source - GPL-3.0 licensed, community-driven

Comparison

| Feature | intelliPrompt | ciga2011's PromptOptimizer | |---------|--------------|---------------------------| | Bug-free | ✅ Yes | ❌ No (unassigned variable bug) | | Offline mode | ✅ Yes | ❌ No | | Fallback on error | ✅ Yes | ❌ No | | Typo correction | ✅ Yes | ❌ No | | Parenthesis balancing | ✅ Yes | ❌ No |

License

GPL-3.0 - Open source and free to use.

Contributing

Contributions are welcome! Please feel free to submit issues or pull requests.

Acknowledgments

• Uses Pollinations for AI-powered prompt optimization
• Inspired by the ComfyUI community's need for better prompt handling

Links

• GitHub Repository
• Report Issues
• ComfyUI