Nodes

Gemini AI Text Generation

Configure the current AI node, powered by Google Gemini.

On this page

Node settings
Available models
System instructions vs prompt
Output
Example workflows
Common AI errors

Node settings

Setting	Purpose
Credential	Choose a saved Google Gemini credential, or `Server environment key` when the deployment provides one.
Variable Name	Names this node's result. Default: `aiText`.
Gemini Model	Chooses one model from the current repository model list.
System Instructions	Optional overall guidance for tone, role, and constraints.
Prompt	Required task/input. Templates are resolved before the provider call.

Available models

Gemini 3.5 Flash-Lite (gemini-3.5-flash-lite) — current default
Gemini 3.6 Flash (gemini-3.6-flash)
Gemini 3.5 Flash (gemini-3.5-flash)

Note

This list reflects the current repository. If Gemini rejects a model as unavailable, reopen the node and choose another listed model.

System instructions vs prompt

System Instructions describe how the model should behave. Prompt describes the specific work for this execution and often includes upstream data.

text

System Instructions:
You are a concise and helpful assistant.

Prompt:
{{discord.command.options.message}}

Use later:
{{aiText.text}}

Output

json

{
  "text": "Generated response",
  "provider": "google",
  "model": "gemini-3.5-flash-lite",
  "usage": {
    "inputTokens": 12,
    "outputTokens": 24,
    "totalTokens": 36
  },
  "finishReason": "stop"
}

The common downstream template is {{aiText.text}}. Execution details deliberately show only safe provider/model/usage metadata, not generated text.

Example workflows

Discord Trigger

Gemini

Send Discord Reply

Telegram Trigger

Gemini

Send Telegram Message

Google Forms Trigger

Gemini

Send Telegram Message or HTTP Request

Common AI errors

Message	Action
Prompt is required	Enter a non-empty prompt and save the node.
Selected Gemini credential is unavailable	Choose an existing owned Gemini credential, or verify the deployment's server key option.
Credential encryption is not configured	The deployment administrator must configure credential encryption; users cannot fix this in the node.
Gemini rejected the selected API key	Replace or correct the Gemini credential.
Gemini quota exceeded	Retry later or review the key's Google plan and billing.
Selected Gemini model is unavailable	Choose a current model from the node settings.

Node settings

Setting	Purpose
Credential	Choose a saved Google Gemini credential, or `Server environment key` when the deployment provides one.
Variable Name	Names this node's result. Default: `aiText`.
Gemini Model	Chooses one model from the current repository model list.
System Instructions	Optional overall guidance for tone, role, and constraints.
Prompt	Required task/input. Templates are resolved before the provider call.

Output

json

{
  "text": "Generated response",
  "provider": "google",
  "model": "gemini-3.5-flash-lite",
  "usage": {
    "inputTokens": 12,
    "outputTokens": 24,
    "totalTokens": 36
  },
  "finishReason": "stop"
}

The common downstream template is {{aiText.text}}. Execution details deliberately show only safe provider/model/usage metadata, not generated text.

Common AI errors

Message	Action
Prompt is required	Enter a non-empty prompt and save the node.
Selected Gemini credential is unavailable	Choose an existing owned Gemini credential, or verify the deployment's server key option.
Credential encryption is not configured	The deployment administrator must configure credential encryption; users cannot fix this in the node.
Gemini rejected the selected API key	Replace or correct the Gemini credential.
Gemini quota exceeded	Retry later or review the key's Google plan and billing.
Selected Gemini model is unavailable	Choose a current model from the node settings.