Run Pipeline
Run a multi-step pipeline on your JSON data directly from the editor.
What pipelines do
A pipeline chains multiple utilities together in sequence:
- Each step takes the output of the previous step as input
- Steps can use different utilities with different configurations
- The final output is the result of the last step
This lets you build complex transformations without running utilities one at a time.
If you want help creating that first draft, start with AI Draft. It can generate a pipeline from plain-language instructions, then hand it off for further review or editing.
How to run a pipeline
1. Open the Command Palette
Press Ctrl+K (or Cmd+K on Mac).
2. Search for a pipeline
Type the pipeline name.
Pipelines appear under the Pipelines section in the results.
3. Review and run
The pipeline modal opens with three panels:
Inputs (left)
- Shows the input data for the pipeline
- If the pipeline has multiple input nodes, switch between them with the dropdown
Pipeline (middle)
- Shows a summary of all steps in order
- Each step displays its name and the utility it uses
Output (right)
- Empty until you click RUN
- Shows the final result after all steps complete
- Displays total execution time
4. Apply the result
- Click RUN to execute all steps
- Review the final output
- Click APPLY to replace your panel content with the result
- Or click Cancel to discard
Things to know
- Steps execute in order — if a step fails, execution stops and the error is shown
- If the pipeline has no input nodes, a message is displayed
- If no final output is produced, an error message explains why
- File size limits apply to the output
Pipelines vs utilities
| Utility | Pipeline | |
|---|---|---|
| Steps | Single operation | Multiple chained operations |
| Configuration | Per-utility settings | Per-step settings (configured in pipeline editor) |
| Use case | Quick one-off transformation | Complex multi-step workflows |
Related
Last updated on