🔍 ComfyUI Metadata Extractor

A powerful ComfyUI custom node that extracts workflow metadata from PNG images and video files generated by ComfyUI. Never lose your prompts, settings, or generation parameters again!

✨ Features

• 📸 Dual Format Support: Extract metadata from both PNG images and video files (MP4, AVI, MOV, MKV, WebM)
• 🎯 9 Output Types: Access all metadata fields individually for maximum flexibility
• 🔄 Multiple Input Methods: File dropdown, manual path entry, or video tensor input
• 🎬 Video Metadata: Full support for ComfyUI-generated video files with embedded workflows
• 📊 Comprehensive Data: Extract prompts, workflows, models, parameters, and more
• 🖥️ Console Output: View formatted metadata directly in your console
• ⚡ Easy Integration: Works seamlessly with existing ComfyUI workflows

📸 Screenshot

Example workflow showing the Metadata Extractor node in action

📦 Installation

Method 1: ComfyUI Manager (Recommended)

1. Open ComfyUI Manager
2. Search for "Metadata Extractor"
3. Click Install
4. Restart ComfyUI

Method 2: Manual Installation

cd ComfyUI/custom_nodes
git clone https://github.com/ai-joe-git/ComfyUI-Metadata-Extractor.git
cd ComfyUI-Metadata-Extractor
pip install -r requirements.txt

Method 3: ffmpeg Installation (Required for Video Support)

Windows:

# Using Chocolatey
choco install ffmpeg

# Or download from: https://ffmpeg.org/download.html

Linux:

sudo apt-get install ffmpeg
# or
sudo yum install ffmpeg

macOS:

brew install ffmpeg

🚀 Usage

Input Options

The node provides three ways to specify your file:

1. File Dropdown: Select from files in your ComfyUI output folder
2. Manual Path: Enter a custom file path directly
3. Video Tensor: Pass a video tensor from another node (for video files)

Outputs

Output	Type	Description
prompt	STRING	The positive prompt used for generation
workflow	STRING	Complete workflow JSON
workflow_json	STRING	Parsed workflow as JSON object
all_text	STRING	All text fields combined
parameters	STRING	Generation parameters (steps, CFG, sampler, etc.)
model	STRING	Model name used
positive	STRING	Positive prompt text
negative	STRING	Negative prompt text
metadata_dict	DICT	All metadata as Python dictionary

Example Workflow Structure

Load Image/Video
    ↓
Metadata Extractor
    ↓
├─ Use prompt output
├─ Use workflow output
├─ Use parameters output
└─ Use model output

💡 Use Cases

• Recreate Past Generations: Extract exact settings from your best outputs
• Organize Your Work: Automatically catalog prompts and settings
• Study Successful Prompts: Analyze what worked in previous generations
• Batch Analysis: Extract metadata from multiple files for comparison
• Workflow Recovery: Restore lost workflows from output images
• Video Metadata: Access generation data from ComfyUI video outputs

📋 Supported Formats

Images

• PNG (with embedded ComfyUI metadata)

Videos

• MP4
• AVI
• MOV
• MKV
• WebM

🔧 Technical Details

Metadata Sources

The node extracts metadata from multiple sources:

1. PNG tEXt chunks: Standard ComfyUI metadata storage
2. Video file metadata: Embedded in video container metadata
3. Workflow JSON: Complete node graph and connections
4. Parameters: Extracted from workflow nodes

Node Detection

The extractor automatically identifies common ComfyUI nodes:

• KSampler
• KSamplerAdvanced
• CLIPTextEncode (positive/negative prompts)
• CheckpointLoaderSimple (model names)
• And more...

🐛 Troubleshooting

"ffprobe not found" Error

Install ffmpeg using the instructions in the Installation section above.

"No metadata found" Error

• Ensure the file was generated by ComfyUI
• Check that the file contains embedded metadata
• For PNG files, verify the image has ComfyUI workflow data
• For video files, ensure they were rendered with metadata enabled

Empty Outputs

If specific outputs are empty:

• The metadata field may not exist in the source file
• Check the console output to see available metadata
• Verify the workflow used nodes that store that information

Video Tensor Input Not Working

• Ensure you're passing a valid video tensor
• Check that the video format is supported
• Verify ffmpeg is properly installed

📝 Console Output

When the node runs, it prints detailed metadata to the console:

=== METADATA EXTRACTION ===
File: output_video.mp4
Prompt: a beautiful landscape...
Model: sd_xl_base_1.0.safetensors
Parameters: steps: 20, cfg: 7.0, sampler: euler_ancestral
===========================

🤝 Contributing

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

📄 License

MIT License - feel free to use and modify as needed.

🔗 Links

• GitHub Repository
• ComfyUI
• Report Issues

⭐ Support

If you find this node useful, please consider:

• ⭐ Starring the repository
• 🐛 Reporting bugs
• 💡 Suggesting new features
• 🤝 Contributing improvements

---

Made with ❤️ for the ComfyUI community