Embeddings Google Gemini Extended icon

Embeddings Google Gemini Extended

Use Google Gemini Embeddings with extended features like output dimensions support

Embeddings Google Gemini Extended

AI-generated

Summary

Provides a Google Gemini embeddings model to connected AI nodes, embedding document text with configurable task type, batch size, newline handling, and output dimensions.

Inputs

  • Model Name — Gemini embedding model to use; defaults to text-embedding-004.
  • Output Dimensions — Requested embedding dimensionality. Set to 0 (the default) to use the model default; only certain models support this setting.
  • Options → Task Type — Embedding use case, such as Retrieval Document (default), Retrieval Query, Semantic Similarity, or Classification.
  • Options → Title — Optional title sent with the text when Task Type is Retrieval Document.
  • Options → Strip New Lines — Replaces newline characters with spaces before embedding. Defaults to true.
  • Options → Batch Size — Maximum texts per request for models that support batching; defaults to 100. The gemini-embedding-001 model is processed one text at a time.

Output shape

An AI embeddings object supplied to connected nodes, with methods for embedding documents.

Requires Google Palm API credentials. Embedding requests are sent to the Gemini API; API errors are thrown. The node is a sub-node with no regular input or output items.

Links

Discussion