Artificial intelligence script steps

Artificial intelligence (AI) script steps allow you to work with large language models (LLMs) and Core ML models. With these steps you can:

• configure a named AI account to use in AI script steps and functions

• get a text response from an AI model given a user prompt

• get a result from your database by sending a model a natural language prompt and your database schema

• get a text response from a model that, given a natural language prompt and database schema, returns an SQL query to use to get a result from the database

• get a FileMaker find request and a found set of records by sending a model a natural language prompt and the fields on the current layout

• insert the embedding vector representation of data into a field in a record or all records in a found set

• perform a semantic find for records with field data that's similar in meaning to a query

• set up a prompt template to use in other AI script steps

• configure a regression model for use

• configure a named RAG account and use it to add or remove data and send prompts to a RAG space

• fine-tune a model with a training data set

• control the logging of AI calls

• configure a Core ML model for use