Gemini step

step_type: gemini

Generates text with Google Gemini directly inside a workflow — write a prompt, optionally set a system instruction, and reference workflow data via expressions.

Configuration

Field	Type	Required	Description
api_key	string	optional	Gemini API key. Use a workspace secret, e.g. secrets.GEMINI_KEY, rather than a literal key.
prompt	string	required	The user message sent to Gemini. Supports expressions — reference trigger or step output data.
system	string	optional	Optional system instruction to set the assistant's tone, role, or output format.
model	enum	optional	Gemini model to use: gemini-2.0-flash (default), gemini-2.0-flash-lite, gemini-2.5-flash-preview-05-20, or gemini-2.5-pro-preview-06-05.
temperature	number	optional	Controls randomness, from 0 (deterministic) to 2 (very creative). Default 1.0.
max_tokens	integer	optional	Maximum tokens in the generated response, from 1 to 8192. Default 1024.

When to use it

• Summarizing an order, product description, or support message mid-workflow.
• Classifying or tagging incoming data before branching with a Condition step.
• Drafting content — email copy, product descriptions — from structured trigger data.

Note

prompt is expression-aware, so it can pull in live values like trigger.payload.title; system is treated as static text.