Merge
Combine several upstream values into one payload for the next step.
Audience
Creators building and publishing workflows.
Overview#
Merge joins multiple data streams before a single downstream block runs. It is the compact way to feed several answers or branches into one prompt, model call, or output.
Common pattern: several Workflow Input blocks → Merge → one LLM Chat block.
Merge
Ports#
- Inputs:
in-1throughin-10: The canvas shows the configured number of numbered inputs (three by default). Each port accepts one connection. - Output:
out(Data): The combined value passed downstream. Older saved handles are normalized at load and execution boundaries.
Inspector#
Choose 2 to 10 visible input ports and one mode: concatenate text, ordered array, shallow object merge, or first non-empty. Existing Merge blocks default to three inputs and text mode.
Tips#
- Connect inputs that belong together semantically, not unrelated leftovers.
- If the graph is hard to read, split merging into stages instead of one crowded merge.
- Use Merge when several cues should feed one prompt, not as a catch-all for messy wiring.
- Label upstream blocks clearly so merged data stays easy to reason about.
Related#
Was this useful?
Your response helps us improve the documentation.