Query Embedding Generator icon

Query Embedding Generator

Generate embeddings for text queries using Vertex AI

Query Embedding Generator

AI-generated

Summary

Generate a numeric embedding vector for a given text query by calling Google Cloud's Vertex AI embedding models, returning the embedding and its dimension with the original input.

Inputs

  • Project ID (required) — The Google Cloud Project ID where the embedding model is hosted.
  • Location (required) — The Google Cloud region where Vertex AI services are available.
  • Query Text (required) — The text string to convert into an embedding vector.
  • Embedding Model (required) — Select the Vertex AI embedding model to use (choices: Text Embedding Gecko 003, 002, or Multilingual Gecko).

Output shape

a single item with JSON containing the original input fields plus 'queryText', an 'embedding' array of numbers representing the vector, and 'embeddingDimension' indicating the vector length.

If the embedding generation fails for an input item, an error message is returned for that item if 'Continue On Fail' is enabled; otherwise, the node stops with an error. The embedding dimension corresponds to the selected model (768 dimensions). The output pairs back to the input item index.

Discussion