Save To Zip¶
Collects one or more connected node outputs, organizes them into a structured temporary folder, and bundles everything into a single zip archive. It auto-names folders/files based on the source nodes and any provided metadata, and returns a link to access the generated archive.
Usage¶
Use this node at the end of a workflow to export results as a single downloadable archive. Connect any node outputs to results_1, results_2, etc. Once connected, the next results_N input becomes available, allowing you to aggregate many outputs. Then connect the produced url to an Output node to click and download the archive.
Inputs¶
| Field | Required | Type | Description | Example |
|---|---|---|---|---|
| results_1...results_101 | False | WILDCARD | Accepts any output from any node. Each connected input is saved into the archive. The next results_N input becomes visible after connecting the previous one. | Any data type (e.g., strings, dicts, model outputs) |
Outputs¶
| Field | Type | Description | Example |
|---|---|---|---|
| url | STRING | A link to the created archive location. Connect this to an Output node to click and download locally. | https://server/output/WORKFLOWID_0/WORKFLOWID_0.zip |
Important Notes¶
- Dynamic inputs: results_1 is required; results_2 through results_101 appear progressively as you connect the previous one.
- Archiving behavior: All connected inputs are written into a folder structure and zipped. JSON-like content is saved as formatted JSON when possible; other types are stringified.
- Metadata-aware saving: If an input carries Salt metadata (folder_name, file_name, seed, config), it will be placed accordingly, and seeds.json/configs.json may be added per folder.
- Batch dictionaries: For certain structured inputs like PDB or A3M dictionaries (when metadata indicates), items are split into per-entry files inside a subfolder.
- Organization for non-metadata inputs: Inputs without metadata are grouped by their source node’s class into uniquely named folders; file names follow output_
.txt. - Execution priority: This node acts as an output node to finalize and produce the archive.
- Unique naming per run: The archive is uniquely identified using workflow_execution_id and the node’s position in the workflow.
- Download tip: The url should be connected to an Output node for clickable download access.
Troubleshooting¶
- No download link appears: Ensure you connected the url output of this node to an Output node to expose a clickable link.
- Missing files in archive: Verify the expected inputs are connected to results_N. Only connected inputs are included.
- Unexpected folder/file names: If inputs lack metadata, the node names folders based on the source node’s class and output index; add metadata-aware nodes upstream if you need precise names.
- Incorrect content format: Non-JSON data is saved as text via string conversion. If you require structured JSON, ensure upstream nodes provide JSON-serializable data.
- Archive collision concerns: workflow_execution_id ensures uniqueness. If multiple Save To Zip nodes exist, each gets a unique index; confirm the correct link is used.
Example Pipelines¶
