Embeddings Block

The Embeddings block converts text into numerical vector representations using OpenAI's embedding models. Reach for it when you need to turn text into embeddings for semantic search, clustering, recommendations, or any other vector-based operation.

Overview

When to Use

• Build semantic search by embedding documents and queries for similarity comparison
• Power retrieval-augmented generation (RAG) pipelines that need vectorized text
• Cluster or deduplicate text by comparing embedding vectors
• Generate features for classification or recommendation systems
• Index content into a vector store for later nearest-neighbor lookup

Configuration

Input Text

The text to convert into an embedding. Required. Typically references an upstream block's output, e.g. {{start.input}}.

Model

Dropdown to choose the OpenAI embedding model. Options: text-embedding-3-small (default), text-embedding-3-large, and text-embedding-ada-002.

API Key

Your OpenAI API key. Required and stored as a password field.

Inputs & Outputs

• Inputs:

  • input (string) — Text to embed
  • model (string) — Embedding model to use
  • apiKey (string) — OpenAI API key

• Outputs:

  • embeddings (json) — Generated embeddings (array of float vectors)
  • model (string) — Model used for generating embeddings
  • usage (json) — Token usage information (prompt_tokens, total_tokens)

Tools

• OpenAI Embeddings (openai_embeddings) — Calls POST https://api.openai.com/v1/embeddings with the given text and model. Returns an array of embedding vectors plus token-usage metadata.

YAML Example

embeddings_1:
  type: openai
  name: "Embeddings"
  inputs:
    input: {{start.input}}
    model: "text-embedding-3-small"
    apiKey: "{{OPENAI_API_KEY}}"
  connections:
    outgoing:
      - target: next-block-id