Vertex AI Advanced icon

Vertex AI Advanced

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

Actions9

Video → Analyze Video

AI-generated

Summary

Analyze videos by providing URLs or binary video files using specified Vertex AI models, returning descriptive analyses of their content.

Inputs

  • Project ID (required) — The Google Cloud Platform project ID.
  • Model (required) — Select the Vertex AI model to use for analysis, either from a list of compatible models or by specifying its ID.
  • Text Input — Optional prompt guiding the analysis, defaulting to "What's in this video?".
  • Input Type (required) — Specify whether the video input is one or more URLs (comma-separated) or binary files from previous nodes.
  • URL(s) — Comma-separated URLs of video files to analyze, required if input type is URL.
  • Input Data Field Name(s) — Name(s) of the binary fields containing video data to analyze, required if input type is binary; multiple fields can be comma-separated.
  • Simplify Output — Choose whether to simplify the raw API response into a more concise format.
  • Options — Additional settings such as maximum tokens to limit the length and detail of the video description.

Output shape

a list of analysis results corresponding to each video input, each as a JSON object with descriptive data about the video's content

The output contains one JSON object per video analyzed. If 'Simplify Output' is true, the results are simplified; otherwise, full API response details are included. Multiple URLs or binary fields result in multiple outputs. Errors thrown on failures.

Examples

Example 1: Analyze a single video URL to summarize its content using a chosen Vertex AI model.

Set Project ID, select Model from list, keep default Text Input, choose 'Video URL(s)' as Input Type, provide one video URL, leave Simplify Output enabled.

Example 2: Analyze multiple binary video files uploaded in previous nodes for content analysis.

Set Project ID, specify Model ID directly, adjust Text Input prompt as needed, choose 'Binary File(s)' as Input Type, list binary fields containing videos as 'data' or custom names separated by commas, optionally disable Simplify Output for full details.

Links

Discussion