π Image Save To Cloud
Push your renders straight to S3 or OSS when the run finishes
- images
If you've ever scripted a ComfyUI batch on a rented GPU and then had to scrape outputs back off the box, you know the pain this node exists to kill. π Image Save To Cloud is the upload half of the pack's cloud pair: give it images plus your S3 or Aliyun OSS details, and when the graph runs it saves each image to a temp folder and pushes it to your bucket. No SSH, no downloading from a dashboard, no manual sync - the render just lands in the cloud, ready for the next stage of your pipeline.
How it works
The required inputs mirror its sibling loader: images (the IMAGE batch), cloud (aws-s3 / aliyun-oss), bucket, region, and your access_key_id / access_key_secret. Then there's filepath, and this one's worth reading closely because it's a format string with defaults:
ComfyUI_{0:05d}_{1:%Y-%m-%d_%H:%M:%S}.png
{0} is the batch index (zero-padded to 5 digits) and {1} is the current timestamp, so the default produces names like ComfyUI_00003_2026-08-16_14:30:22.png. You can rework the pattern - change the folder prefix, swap the timestamp format - but keep the {0}/{1} placeholders or every image in a batch gets the same filename.
On execution it converts each image to PNG, writes it to ComfyUI's temp folder, then uploads via boto3 (S3) or oss2 (Aliyun). It's marked as an output node, and notably it has no outputs - it's a terminal step that does its work and shows the uploaded images in the UI.
Inputs and outputs
- All required inputs as above (no optional inputs, no outputs).
- The node is an output node: it runs as a terminal step and displays what it saved.
Where it fits
The obvious pairing is with π Image Load From Cloud for a full cloud round-trip, or as the final step of a headless generation farm that stages results into a shared bucket. It's genuinely useful for anyone running ComfyUI as a service rather than a desktop toy - for everyone else it's a lot of credentials to type for what a Save Image node already does locally.
Install
Part of hay86/ComfyUI_AceNodes:
cd ComfyUI/custom_nodes
git clone https://github.com/hay86/ComfyUI_AceNodes
cd ComfyUI_AceNodes
pip install -r requirements.txt
Restart ComfyUI and look for π
Image Save To Cloud under Ace Nodes. Manager users search ComfyUI_AceNodes.
Gotchas
Same credential warning as the loader: access keys live in plaintext on the node, so treat any saved workflow containing them as a secret and use keys scoped to just that bucket. If the upload fails, the node swallows the SDK error to a console print and the workflow still reports success - so verify objects actually landed in the bucket. And the pack's heavy requirements.txt (rembg, transformers, insightfaceβ¦) installs regardless. Niche, but for the cloud-pipeline crowd, exactly the tool that's usually missing.
Inputs (7)
| Name | Type | Default | Description |
|---|---|---|---|
| images | IMAGE | β | |
| filepath | STRING | ComfyUI_{0:05d}_{1:%Y-%m-%d_%H:%M:%S}.png | β |
| bucket | STRING | β | |
| region | STRING | β | |
| cloud | COMBO | 2 options: aws-s3, aliyun-oss | |
| access_key_id | STRING | β | |
| access_key_secret | STRING | β |
Outputs (0)
No outputs