SplitTextNode, then runs a sub-pipeline once per element by setting execution_mode="batch" on the PipelineNode.
You’ll end up with. A saved batched-pipeline whose output is the sub-pipeline’s output_0 produced for each split line.
Expected output
execution_mode="batch" is what tells the runtime to fan the sub-pipeline out across the list — without it, the node receives the list as a single input.
See also
Sub-pipelines
Single-shot sub-pipeline composition.
Batch config
Batch mode on individual LLM nodes.
Pipeline reference
Full method surface.
