Extensions/Comfyui_Attome_S3
ComfyUI Extension

Comfyui_Attome_S3

Custom nodes for ComfyUI enabling seamless loading and saving of media files (images, videos, audio, text) directly from/to AWS S3 or S3-compatible storage services.

By attome-ai·Created 8 months ago·Updated 7 months ago· 4
attome-ai/Comfyui_Attome_S3
Nodes9
On cloudLocal install
CategoryAttome/S3
Stars4
Updated7 months ago

Nodes (9)

Attome S3 Config
Attome/S3
Attome S3 Load Audio
Attome/S3
Attome S3 Load Image
Attome/S3
Attome S3 Load Text
Attome/S3
Attome S3 Load Video
Attome/S3
Attome S3 Save Audio
Attome/S3
Attome S3 Save Image
Attome/S3
Attome S3 Save Text
Attome/S3
Attome S3 Save Video
Attome/S3
Readme

ComfyUI Attome S3 Nodes

Custom nodes for ComfyUI that enable seamless loading and saving of media files (images, videos, audio, text) directly from/to AWS S3 or S3-compatible storage services.

ComfyUI Python License

Features

  • Optional S3 Config - Use env.txt defaults or override with config node when needed
  • Environment Configuration - Load default S3 credentials from env.txt file (cached for performance)
  • Smart S3 Key Handling - Returns proper empty resources (512×512 images, 1024-sample audio) when s3_key is empty
  • Automatic File Extensions - File extensions auto-match selected format (PNG/JPEG/WEBP, WAV/MP3, etc.)
  • Metadata Control - Optional workflow metadata embedding in images
  • Batch Saving - Save multiple images/files at once with automatic numbering (output_1.png, output_2.png, etc.)
  • Dynamic Filenames - Use placeholders like %date%, %time%, %random% in your file paths
  • List Output - All save nodes return a list of S3 URIs for easy downstream processing
  • Image Operations - Load/Save PNG, JPEG, WEBP with quality settings
  • Video Operations - Load/Save MP4 with frame extraction options
  • Audio Operations - Load/Save WAV, MP3, FLAC, OGG formats
  • Text Operations - Load/Save text files with encoding support
  • S3-Compatible - Works with AWS S3, Rustfs, MinIO, DigitalOcean Spaces, Backblaze B2, etc.

Nodes

| Node | Description | |------|-------------| | Attome S3 Config | Configure AWS credentials, region, bucket, and optional endpoint URL | | Attome S3 Load Image | Load image from S3, outputs IMAGE + MASK tensors | | Attome S3 Save Image | Save image to S3 (PNG/JPEG/WEBP, configurable quality) | | Attome S3 Load Video | Load video from S3, outputs frames as IMAGE batch + FPS | | Attome S3 Save Video | Save image batch as video to S3 (configurable FPS/codec) | | Attome S3 Load Audio | Load audio from S3, outputs AUDIO dict | | Attome S3 Save Audio | Save audio to S3 (WAV/MP3/FLAC/OGG) | | Attome S3 Load Text | Load text file from S3 | | Attome S3 Save Text | Save text content to S3 |

<img width="1016" height="748" alt="image" src="https://github.com/user-attachments/assets/0aa82b3c-2556-4894-a65d-05765bbcc6d1" />

Installation

Method 1: Portable ComfyUI (Windows)

cd ComfyUI_windows_portable\ComfyUI\custom_nodes
git clone https://github.com/attome-ai/Comfyui_Attome_S3.git
cd Comfyui_Attome_S3
..\..\..\python_embeded\python.exe -m pip install -r requirements.txt

Requirements

  • Python 3.10+
  • ComfyUI
  • boto3
  • Pillow (usually pre-installed with ComfyUI)
  • torch (usually pre-installed with ComfyUI)
  • numpy (usually pre-installed with ComfyUI)
  • torchaudio (for audio operations)
  • opencv-python (for video operations)

Usage

Basic Workflow

Option 1: Use env.txt (Recommended for single environment)

  1. Edit env.txt with your credentials (see Using env.txt)
  2. Use any Load/Save node directly - no config node needed!
  3. Specify the S3 key (path) for your file

Option 2: Use Config Node (For multiple environments or overrides)

  1. Add Attome S3 Config node
  2. Enter your AWS credentials:
    • aws_access_key_id: Your AWS access key
    • aws_secret_access_key: Your AWS secret key
    • region_name: AWS region (e.g., us-east-1)
    • bucket_name: Your S3 bucket name
    • endpoint_url: (Optional) Custom endpoint for S3-compatible services
  3. Connect the s3_config output to any Load/Save node
  4. Specify the S3 key (path) for your file

Example: Load Image from S3

Without Config Node (using env.txt):

[Attome S3 Load Image] --> [Your Workflow]
         |
         +-- s3_key: "images/input.png"

With Config Node:

[Attome S3 Config] --> [Attome S3 Load Image] --> [Your Workflow]
                             |
                             +-- s3_key: "images/input.png"

Example: Save Generated Image to S3

[Your Workflow] --> [Attome S3 Save Image] <-- [Attome S3 Config]
                            |
                            +-- s3_key: "outputs/result.png"
                            +-- format: "PNG"
                            +-- quality: 95
                            +-- save_metadata: true (or false)

Workflow Metadata Control

The save_metadata parameter controls whether ComfyUI workflow information is embedded in saved files:

✅ save_metadata: true (Default)

  • Embeds complete workflow data (nodes, connections, parameters) in file metadata
  • Allows recovery of the exact workflow that generated the file
  • Perfect for debugging and testing workflows
  • Slightly larger file size due to embedded JSON

❌ save_metadata: false

  • Saves clean files without any workflow information
  • Smaller file sizes
  • No exposed workflow data when sharing publicly
  • Perfect for production outputs

Format-Specific Implementation:

| Format | Metadata Storage Method | Notes | |--------|------------------------|-------| | PNG Images | PNG text chunks (pnginfo) | Standard method, works with all PNG viewers | | MP4 Videos | MP4 metadata tags (©cmt, desc) | Requires mutagen library (optional) | | MP3/FLAC Audio | ID3/Vorbis comment tags | Requires mutagen library (optional) | | Text Files | Comment header (# ComfyUI Workflow...) | Prepended as comments at file start | | JPEG/WEBP/WAV/OGG | Not supported | Metadata will be silently skipped |

Use Cases:

  • Development/Testing: Keep save_metadata: true to easily recover and debug workflows
  • Production/Sharing: Set save_metadata: false for clean, compact files
  • Public APIs: Disable to prevent exposing your workflow logic

Optional Dependency:

pip install mutagen  # For video/audio metadata support

Using with S3-Compatible Services

For MinIO, DigitalOcean Spaces, Backblaze B2, or other S3-compatible services, set the endpoint_url parameter:

| Service | Endpoint URL Example | |---------|---------------------| | MinIO | http://localhost:9000 | | DigitalOcean Spaces | https://nyc3.digitaloceanspaces.com | | Backblaze B2 | https://s3.us-west-000.backblazeb2.com | | Cloudflare R2 | https://<account_id>.r2.cloudflarestorage.com |

Using env.txt for Default Configuration

You can use the env.txt file in the plugin directory to set default values for the S3 Config node. This is useful for avoiding repeated credential entry.

  1. Edit env.txt in the plugin directory with your credentials:
AWS_ACCESS_KEY_ID=your_access_key_here
AWS_SECRET_ACCESS_KEY=your_secret_key_here
REGION_NAME=us-east-1
BUCKET_NAME=your_bucket_name_here
ENDPOINT_URL=
  1. Restart ComfyUI - all S3 nodes will now use these as defaults automatically (no config node required!)

Benefits:

  • ✅ No need to wire config nodes in every workflow
  • ✅ Cached in memory - only reads file once per session
  • ✅ Can still override with config node when needed

Empty S3 Key Behavior

All load nodes intelligently handle empty s3_key values:

  • Empty Image: Returns 512×512 black image (prevents division errors in nodes like IPAdapter)
  • Empty Video: Returns single 512×512 black frame
  • Empty Audio: Returns 1024-sample silent audio
  • Empty Text: Returns empty string

This allows optional S3 resources in workflows without errors.

Automatic File Extension Matching

Save nodes automatically adjust file extensions to match your format selection:

Path: "output"  + Format: WEBP  = Saved as: "output.webp"
Path: "output.png"  + Format: JPEG  = Saved as: "output.jpg"

Supported Formats:

  • Images: PNG, JPEG (.jpg), WEBP
  • Audio: WAV, MP3, FLAC, OGG
  • Video: MP4 (always)

Troubleshooting

"No module named 'boto3'"

pip install boto3

"No module named 'cv2'" (Video nodes)

pip install opencv-python

"No module named 'torchaudio'" (Audio nodes)

pip install torchaudio

"Access Denied" errors

  1. Verify your AWS credentials are correct
  2. Check the IAM user has the required S3 permissions
  3. Ensure the bucket name and region are correct
  4. For S3-compatible services, verify the endpoint URL

Nodes not appearing in ComfyUI

  1. Restart ComfyUI completely
  2. Check the console for import errors
  3. Verify all dependencies are installed

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

  1. Fork the repository
  2. Create your feature branch (git checkout -b feature/AmazingFeature)
  3. Commit your changes (git commit -m 'Add some AmazingFeature')
  4. Push to the branch (git push origin feature/AmazingFeature)
  5. Open a Pull Request

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

  • ComfyUI - The amazing node-based Stable Diffusion UI
  • boto3 - AWS SDK for Python

Support

If you encounter any issues or have questions:

  1. Check the Troubleshooting section
  2. Search existing issues
  3. Open a new issue with:
    • ComfyUI version
    • Python version
    • Full error message
    • Steps to reproduce