@seekdb/google-vertex

Google Vertex AI embedding function for SeekDB.

Google Vertex AI provides text embedding models. seekdb provides a GoogleVertexEmbeddingFunction wrapper so you can generate embeddings via Vertex AI and use them with seekdb collections.

Authentication (ADC)

GoogleVertexEmbeddingFunction calls the Vertex AI embedding API via the Google Cloud SDK. In practice, you typically need:

• A Google Cloud project with Vertex AI enabled and permission to invoke embedding models
• Google Application Default Credentials (ADC) configured for your environment

Installation

npm i seekdb @seekdb/google-vertex

Usage

import { GoogleVertexEmbeddingFunction } from "@seekdb/google-vertex";

const ef = new GoogleVertexEmbeddingFunction({
  projectId: "your-gcp-project-id",
  modelName: "textembedding-gecko",
});

Configuration

• projectId: GCP project id (required)
• location: Vertex AI location (default: "us-central1")
• modelName: model name (default: "textembedding-gecko")