Steps for building AI-powered responses in Claris FileMaker.

Learn how to quickly set up a basic AI-powered feature that interacts with a large language model (LLM). What you build can become the foundation for more complex AI implementations and explorations.

See it in action

Let's get you started:

Once you have these two requirements ready to go, follow these steps:

01

Configure your AI account.

Start by setting up an account with a supported AI provider, such as OpenAI, Anthropic, or Cohere. Use the Configure AI Account script step and provide a name for your account, select a provider — like Open AI — then enter your API key.

02

Add the Generate Response from Model script step.

It's the key step that enables communication with the LLM. Include this script step in your script, specifying the account name you configured earlier.

03

Choose the model.

Select the LLM you want to interact with, such as GPT-40.

04

Provide the prompt.

Determine what message or input you want to send to the model. Use a field or a variable to store the prompt.

05

Set the response field.

Decide where you want to store the response from the model, using a field or a variable to capture the model's output.

06

Turn on streaming — an optional step.

If you want to see the response as it's generated, then turn on streaming. This provides a real-time view of the model's output.

07

Save message history — an optional step.

To create a chatbot-like experience, enable the Save Message History option. This allows you to store the chat history in a variable, providing context for the model's responses. Specify the variable to save the history and the number of previous messages to include.

08

Test and interact.

Run your script and start interacting with the LLM. Provide prompts and see the model's responses. With the message history saved, the model has context for your conversations.

09

Enhance presentation — an optional step.

You can use calculations or web viewers to enhance the presentation of the chatbot interface. Format the response text or use web technologies to make everything more visually appealing.

Take your app further with some experimentation.

Once you've added the new feature to an app, experiment with different prompts, models, and presentation options. This script step opens up a range of possibilities for integrating LLMs directly into your custom apps.

Get expert help if you need it.

Want to discuss additional AI opportunities with a development expert? Let us help set up that conversation. Complete this form.

Prefer a PDF of these instructions?

Download PDF