Embedding

Vector representation of a piece of content, used for similarity (semantic) search

type Embedding {
  content: String
  created_at: JSONDate!
  embedding: [Float!]!
  id: ID!
  metadata: HashObject
  token_count: Int
  updated_at: JSONDate!
}

Fields

content (String)

Original text that was converted to the embedding

created_at (JSONDate)

Timestamp of when the object was created, set automatically by the system

embedding (Float)

Vector with 1536 dimensions, compatible with OpenAI embeddings

id (ID)

Unique identifier of the object

metadata (HashObject)

Arbitrary data stored with the embedding, e.g. a reference to the source record; can be used for filtering

token_count (Int)

Number of tokens of the content, as provided when the embedding was saved

updated_at (JSONDate)

Timestamp of when the object was last updated, set automatically by the system