Embeddings and Perform Semantic Find

Intermediate

Generate embedding vectors, store them efficiently, and search text or images by meaning.

What you'll learn

When to use Insert Embedding vs Insert Embedding in Found Set
Why container fields are preferred for stored embedding vectors
The three Query by modes in Perform Semantic Find

Embeddings turn text or images into vectors. FileMaker can store those vectors in text or container fields, then Perform Semantic Find can compare a query vector to the stored vectors and return the most similar records.

Stuck is a valid status

Need a second brain on this one?

If this lesson just collided with your real schema, script stack, or deadline, book consulting and turn the confusion into a concrete plan.

Book consulting

1/3

Create vectors for one record or a found set

Insert Embedding sends one expression to an embedding model and stores the returned vector in a field or variable. Insert Embedding in Found Set repeats that process for each record in the found set using a source field and target field.

FileMaker Script

Insert Embedding [
  Account Name: "main-ai" ;
  Embedding Model: "text-embedding-3-small" ;
  Input: Articles::Body ;
  Target: Articles::Body_Embedding
]

Insert Embedding in Found Set [
  Account Name: "main-ai" ;
  Embedding Model: "text-embedding-3-small" ;
  Source Field: Articles::Body ;
  Target Field: Articles::Body_Embedding ;
  Replace target contents
]

Perform SQL Query by Natural Language AI functions for vectors, schema, and tokens