This ComfyUI tutorial workflow automates running a list of text prompts one-by-one and saving an image for each entry. It uses the custom FL_PromptSelectorBasic node to read a multiline prompt list and select a single line based on an index provided by a PrimitiveInt node. That selected prompt is then fed into the generation chain powered by the Z-Image-Turbo model, and the resulting image is written to disk by SaveImage. A MarkdownNote in the graph documents the intent and wiring so you can quickly see how the pieces fit together.

Under the hood, FL_PromptSelectorBasic exposes a text field where you paste prompts (one per line). The PrimitiveInt node supplies an index that determines which line is currently active. In a batch run, the index advances one step per batch iteration (when driven by your generation/driver node; listed here as ca7506c4-0f1a-4a77-9d2e-b9077eaf58d5), so each iteration consumes the next prompt from the list. This pattern is ideal for prompt A/B testing, producing sets of assets, or iterating through a campaign brief without manually editing the graph between runs.