What is an embedding?
An embedding is the translation of a text into a long series of numbers that represents its meaning, so that computers can find similar content.

With embeddings, every text gets an address on a huge map of meaning. Texts about holidays live close together, texts about invoices elsewhere. That way I find what fits, even when different words are used.
An everyday example
A customer searches the help section for money back. Thanks to embeddings, the search finds the article about refunds, even though the word money does not appear in it.
What matters
Embeddings are the basis for RAG and for search functions that understand meaning instead of just keywords. They are calculated once and stored in a vector database. If your documents change, they have to be recalculated.
Read on
Related pages
Questions
Rather talk it through?.