# Parallel

Process data in multiple branches, then continue after they join.

Use it for independent operations on the same input, such as text processing and field extraction.

## Organize branches

Configure each branch's nodes or processing chain, including its inputs and results. Steps that depend on one another should run sequentially in the same chain.

The workflow continues at the join after all branches finish. Add a data-processing step after the join if results need to be combined into a table or object.

## Use joined results

Select the returned fields needed by downstream nodes. For example, retrieve the summary from one branch and extracted fields from another, then construct the data to save.

Branches have no fixed execution order.

## Example: generate a summary and fields

Flow: General Document Parsing → Parallel branches → Join → Save results.

One branch sends text to AI Inference for a summary; another sends documents to Information Extraction. Map each result separately after the join to avoid overwriting fields with identical names.

## Related pages

- [Variables and data flow](../../variables-and-data.md)
- [Run and debug](../../run-debug.md)