What is Embedding?
A fixed-length vector representing the meaning of a piece of text. Similar meanings → close vectors. Generated by a separate embedding model (e.g. text-embedding-3-small) and stored in a vector store for fast semantic search.
What an embedding is
A text embedding is a list of numbers that represents the meaning of a piece of text, so that texts with similar meaning end up close together. Anthropic's documentation describes embeddings as numerical representations of text used to measure semantic similarity for tasks such as search, recommendations and anomaly detection.
Where they come from
Anthropic states that it does not offer its own embedding model and points to Voyage AI as one provider. Whichever vendor you pick, the usual factors are how well the model fits your domain, its speed and cost at scale, and whether it can be customised.
How they are used
For retrieval, you embed each document once and each query at search time, then return the documents whose vectors are nearest to the query's. Voyage's guidance is to say whether an input is a query or a document, because that gives better vectors for retrieval. Its embeddings are normalised, so dot product and cosine similarity give the same ranking.
Example
You embed six paragraphs of notes. The query "When is the earnings call?" is embedded the same way, and the paragraph about the scheduled conference call scores highest even though the two share few words.
Sources
← Back to the full AI agent glossary.