ISO FileMaker Magazine

In the video for this article I take a deep dive into FileMaker 26's new field and table metadata features, and the big picture is: Claris is prepping your databases for AI agents. I show how you can pack your field comments with structured JSON—things like enumerations ("colors must be one of: blue, orange, red")—and pull that out via the Get Table DDL function so AI tools have enough schema context to generate scripts or build equivalent tables in other databases. It builds on functionality that actually came in FileMaker 21, but FileMaker 26 extends it with some serious new knobs.

The meat of the demo is the new Advanced tab in the Define Database dialog. You can now override a field's display name, export name, sort order, and even localize those labels by locale. But there's a gotcha that will bite you if you're not paying attention: the moment any field gets an annotation on that Advanced tab, Get Table DDL stops returning all the fields and only returns the ones you explicitly annotated. The unannotated ones just vanish from the output. So you've got to be intentional—leaving the Advanced tab empty means you get everything, but filling it in means only what you declared. I walk through a neat trick for localization: building a custom function using a Let and a Case statement that swaps the field label based on Get System Locale Elements, so "colors" becomes "couleurs" or "Farben" depending on the user's system locale.

I also uncover a weird rabbit hole in the Claris docs—buried at the very bottom of the Get Table DDL best practices page. If you prefix any field comment with [LLM] and leave the Advanced tab untouched, that field gets included in DDL output. It trims the whitespace and strips the tag before returning the value. It looks like Claris kept this as a backwards-compatible signal from a pre-26 system, and it's effectively dead unless you deliberately use the tag. Long story short: if you're doing AI integration with FileMaker, you now have multiple layers of control over what gets exposed; field comments, DDL annotations, display name overrides, and locale-aware labels—and you just need to be aware of which behaviors supersede which when they conflict.

While some of the excitement around AI deals with using the technology within your FileMaker solution, it's likely that, as a developer, you should be more excited about how AI changes what you do as a developer.

In fact, it's theorized that the occupation of being a developer will be one of the careers most affected by AI. So, what do you do while we wait and see if this is truly the case? Well, you use AI, of course!

If you haven't yet dipped your toes into the world of using AI to compose some of the code you need, then make sure not to get left behind. You ABSOLUTELY need to investigate how AI changes how you develop. In my own personal experience, I've likely saved a ton of time by having AI write my code for me. The advantage of a developer who already knows code and can take advantage of AI is that they are better developers than those just starting out with no background in coding.

Yes, maybe AI results will eventually become nearly perfect, such that dictating a prompt to create a solution requires no existing knowledge of code. However, any existing knowledge makes you a better composing developer than someone with no knowledge at all.

Here's the trick: You need to know how to take advantage of the tools available, and this video is going to provide the introduction you need in order to make things much faster when coding your FileMaker solution.