Gemini Batch

Process multiple Gemini requests in a batch.

Overview

Gemini Batch uses the Google Gemini batchGenerateContent endpoint to process multiple content generation requests in a single synchronous API call. Unlike OpenAI and Anthropic batch APIs which are asynchronous, Gemini batch is synchronous — it sends all requests at once and receives all responses in a single response. Each input item becomes one request in the batch. Individual responses are matched back to their input items.

Category: AI
Tool Name: gemini_batch
Version: 1

Appearance: Icon: gemini | Color: #ffffff

Node Type

Action — processes input items and produces output

Input / Output

Direction	Port(s)
Input	Input
Output	Output, Error

Credentials

This tool requires Google AI credentials. See the Credentials Guide for setup instructions.

Parameters

Parameter	Type	Required	Default	Description
Model	options	No	(current default)	The Gemini model to use for batch requests. Pick a specific version — labels are pinned, not aliased to “latest”.
				Options: the Gemini chat models available to your workspace — pick one from the dropdown.
System Prompt	string	No	—	System instructions applied to all batch requests. Supports expressions.
File URIs	string	No	—	Optional. fileUri value(s) from the Gemini File Upload tool, attached to every batch request. Pass a single URI or a JSON array. Per-item override: include a “fileUris” field on the input item. Gemini File API retains uploads for 48 hours. Supports expressions.
Options	collection	No	{}	Optional generation and output settings — add only the fields you need.
— Temperature	number	No	1	Sampling temperature (0-2). Lower values make output more focused.
— Max Output Tokens	number	No	4096	Maximum number of tokens per response.
— Response Field Name	string	No	batch	Field name in the output JSON where the extracted response text will be placed. Do not use “response” — the raw API response object is always written to that key and would overwrite the extracted text.
Include Input	boolean	No	false	Whether to include the original input item fields in the output.
Max Concurrency	number	No	5	Maximum number of items to process concurrently.

Output Data

Every input item becomes one request in a single batch call, and one output item comes back per input item. The generated text lands on the field named by Response Field Name (batch by default), with the raw API response and the item’s position in the batch beside it. The rest of the input item JSON is dropped unless Include Input is on; binary data on the input item is forwarded unchanged.

{
  "batch": "The generated text for this item",
  "response": "the raw Gemini response object for this request",
  "model": "the model that produced the responses",
  "finishReason": "the API's reason for ending generation",
  "usage": { "promptTokenCount": 128, "candidatesTokenCount": 256 },
  "batchIndex": 0,
  "totalRequests": 25
}

• Each item supplies its own prompt. There is no user-message parameter on this node: the prompt is read from the item’s message, prompt or text field, in that order. Put it there with an upstream Edit Fields node.
• An item carrying none of those three fields is still returned, with the response field set to null, status: "skipped" and a reason saying no message, prompt or text field was found. If no item in the run carries a prompt, every item comes back with status: "empty" instead.
• response always holds the raw Gemini response for that request, so do not set Response Field Name to response — the raw object would overwrite the extracted text.
• batchIndex is the item’s position within the batch and totalRequests is how many requests were sent, which is what to log when reconciling a large run.
• usage is the per-request token accounting, present only when the API reported it.

Reference the result downstream by expression, e.g. {{ $json.batch }}.

Usage Examples

• Process 100 prompts in a single Gemini batch call
• Translate multiple texts using Gemini batch
• Generate summaries for a dataset in one API call
• Bulk classify content using Gemini batch
• Run batch content generation with a system prompt

Example Configuration

Summarize every incoming item — each item must carry its prompt on message, prompt or text:

{
  "type": "gemini_batch",
  "parameters": {
    "systemPrompt": "Summarize the following text in 2-3 sentences.",
    "includeInput": true,
    "maxConcurrency": 10,
    "options": {
      "temperature": 0.4,
      "maxOutputTokens": 256,
      "responseFieldName": "summary"
    }
  }
}

High-throughput classification with deterministic output:

{
  "type": "gemini_batch",
  "parameters": {
    "systemPrompt": "Classify the sentiment of the input as POSITIVE, NEGATIVE, or NEUTRAL. Reply with only the label.",
    "includeInput": false,
    "maxConcurrency": 20,
    "options": {
      "temperature": 0,
      "maxOutputTokens": 16,
      "responseFieldName": "sentiment"
    }
  }
}

Ask a batch of questions against one uploaded document:

{
  "type": "gemini_batch",
  "parameters": {
    "systemPrompt": "Answer each question using only the attached document. Reply in one sentence.",
    "fileUris": "{{ $json.fileUri }}",
    "includeInput": true,
    "options": {
      "responseFieldName": "answer"
    }
  }
}

Error Handling

Mode	Behavior
stop	Halts workflow on first error
continue	Skips failed items, passes successful ones through
errorPort	Routes failed items to Error output port

Tips

Gemini Batch sends multiple content generation requests to the Google Gemini batchGenerateContent endpoint in a single synchronous API call. Use it when you need to process a list of prompts or inputs together and want all results returned at once without managing asynchronous job polling. It produces matched response objects on the main output for each successful input item and routes failures to the error output.