Save To Bucket¶
Saves connected node outputs into a structured folder in a cloud bucket and returns a signed URL to a zip snapshot of that folder. It organizes files by source node and any attached metadata (folder/file names), supports many inputs, and merges results using the workflow execution ID to distinguish runs.
Usage¶
Use this node at the end of a workflow to persist all relevant results to a cloud bucket and obtain a downloadable link. Connect any outputs (texts, structured strings, dictionaries like PDB/A3M, etc.) to results_N. Provide a non-empty folder_path to control where the data is stored in the bucket. The node uploads all files, merges them under a unique execution ID, and returns a signed URL to download a zip of the saved contents.
Inputs¶
| Field | Required | Type | Description | Example |
|---|---|---|---|---|
| results_N (N = 1..101) | False | WILDCARD | Connect any output you want to save (strings, dict-like biotech formats, etc.). Inputs with Salt metadata will be written using their provided folder/file names; others will be stored under a folder named after the source node. For dict-like biotech inputs (e.g., PDB/A3M), items are split into separate files. | A PDB dict like {"protein1": "PDB_CONTENT..."} or a STRING such as "inference complete" |
| folder_path | True | STRING | Destination path inside the cloud bucket. Must be non-empty and should not start or end with '/'. Nested paths are supported using forward slashes. | biotech/runs/alphafold_experiment |
Outputs¶
| Field | Type | Description | Example |
|---|---|---|---|
| url | STRING | Signed URL to a zip file that represents the current content of the destination folder in the bucket. Use it to download the saved results. | https://storage.signed-url.example/download?sig=... |
Important Notes¶
- Folder path must be non-empty and should not include leading or trailing slashes; nested folders are allowed (e.g., project/experiment1).
- If multiple inputs result in the same file name, the node will auto-suffix with _2, _3, etc., to avoid overwriting.
- Inputs carrying Salt metadata control their folder/file names and may also add seeds.json and configs.json at the folder level.
- Dictionary inputs for biotech formats (e.g., PDB, A3M) are saved as multiple files, one per entry.
- This node merges results by workflow execution ID so multiple runs under the same folder_path can coexist and be distinguished.
Troubleshooting¶
- Error: 'Please specify non empty folder path.' — Set a valid non-empty folder_path without leading/trailing slashes.
- No download link returned — Ensure at least one results_N input is connected and that the workflow completed without errors.
- Timeout while uploading — Large numbers of files or very large content may exceed the operation window. Reduce output size or try splitting saves.
- Unexpected file naming or structure — Check whether inputs provided Salt metadata; metadata dictates directory and file names. Without metadata, files are placed in a folder named after the source node.
- Missing expected files for dict-like inputs — Verify the input format. For PDB/A3M, ensure the input is a dictionary of name->content.
Example Pipelines¶
