Vertex AI Advanced icon

Vertex AI Advanced

Interact with Google Vertex AI models — multimodal text, image, audio, and video

Actions9

Image → Analyze Image

AI-generated

Summary

Analyze one or more images using a specified Google Vertex AI model to answer a question or describe the images.

Inputs

  • Project ID (required) — The Google Cloud Platform project ID used for the API request.
  • Model (required) — The Vertex AI model to use for analyzing the image(s), selectable from a searchable list or by specifying the model ID.
  • Text Input — A prompt or question for the model describing what to analyze or inquire about in the image(s). Defaults to "What's in this image?".
  • Input Type (required) — Specifies whether to provide image inputs as URLs or as binary files from previous nodes.
  • URL(s) — Comma-separated list of image URLs to analyze when Input Type is set to URLs.
  • Input Data Field Name(s) — Comma-separated list of binary field names containing image file data to analyze when Input Type is set to binary.
  • Simplify Output — Determines whether to return a simplified response focusing on key data.
  • Options — Additional options such as the maximum number of tokens in the description (affects description length and detail).

Output shape

a list of response candidates from the Vertex AI model, each representing an analysis result of the input images

Supports multiple image inputs either as URLs or binary data fields. The response is an array of candidate analysis results unless Simplify Output is false, in which case the full response object is returned as a single item. The node calls the Vertex AI generateContent API with the given model and input images, using a user-provided question or prompt.

Examples

Example 1: Analyze images from URLs to describe their contents

Set Project ID and select a compatible model ID. Choose Input Type as 'Image URL(s)' and provide one or multiple image URLs separated by commas. Optionally adjust the Text Input prompt and the maximum length of the description. Enable Simplify Output for simplified JSON results.

Example 2: Analyze images provided as binary files from a previous node

Set Project ID and select a compatible model ID. Choose Input Type as 'Binary File(s)' and specify the binary property names containing image data (comma-separated). Customize the Text Input and description length as needed. Enable or disable simplified output depending on the detail required.

Links

Discussion