FileMaker DevCast: Everything Claris FileMaker

This episode of the DevCast looks at how JavaScript, HTML, JSON, Web Viewers, and AI can work together to build more flexible and dynamic FileMaker solutions. A few of our devs walk us through real-world examples, like a multi-image upload and rotation tool and a dynamic dining room and table management interface.The conversation strides into a bigger question: How much do you really need to understand when AI is writing the code for you? Our devs discuss finding the right balance between letting AI handle the heavy lifting and understanding enough of what's happening under the hood to build, test, troubleshoot, and maintain a reliable solution.Multi-Image Upload Blog Pt. 1: https://www.portagebay.com/blog/multi-image-uploading-with-gallery-and-rotation-free-demo/Multi-Image Upload Blog Pt. 2: https://www.portagebay.com/blog/using-webviewers-to-add-functionality-for-field-technicians/

It's been two months since FileMaker 2026 was released and our devs are still discovering. This episode, the team is looking into what the update really means for developers and the new features they’re most excited about: AI-assisted development possibilities, WebDirect improvements, layout enhancements, and more. They also discuss some of the quieter but valuable additions in FileMaker 2026, like improved field-name mapping for imports, new PDF commands, and expanded server-side functionality.A couple of our devs even share hands-on experiments with using AI and XML to rebuild and modify solutions and where the current limitations start to show up with larger files.John's Blog Post: https://www.portagebay.com/blog/can-filemaker-2026-rebuild-a-file-clean-testing-generatedbfile/

In this episode of DevCast, our devs share their firsthand experiences using AI tools like Claude and Gemini for FileMaker development. They spotlight where they excel, where they fall short, and what developers need to watch out for. They cover AI hallucinating FileMaker functions and scripts, misinterpreting FileMaker Server logs, context window limitations, memory persistence, and the importance of verifying AI-generated code before putting it into production. They also discuss practical advice on improving AI-assisted workflows, using skills and agents, keeping conversations focused, building better prompts, and why testing is more important than ever in an AI-driven development process.

In their spare time, John, Xandon, and Jacob have been creating personal web apps. John is creating a healthcare referral portal that easily and efficiently collects and processes referral submissions. Xandon is prepping for his kid's college days by having all the info they need, like planning, admissions, and athletic recruiting in one space. Jacob is making a homeschool learning platform for his daughter that helps him design personalized education plans for special-needs students.They share the ups and downs of their processes and lessons they've learned about full-stack development, SaaS products, subscriptions, AI-assisted coding, and turning personal challenges into practical software solutions.

One little ExecuteSQL query can’t hurt… right? In this episode, the team unpacks the FileMaker performance issues hiding inside real-world solutions. Runaway SQL calls, invisible object calculations, and AI-generated code that looked smart until it brought everything to a crawl. They discuss debugging wins, optimization tips, and the kinds of “how did this even happen?” moments that every developer eventually runs into.

Welcome to an inside look at how our devs are testing out where AI can help versus hinder every day work conundrums. Our team's focus in this meeting is Claude Cowork, a tool designed to track everything within a specified local folder.As for team demonstrations and discussions, Xandon uses a plugin that parses FileMaker XML to generate web-based code reviews and actionable recommendations. Nick showcases a clever approach to token efficiency, using Python scripts to convert verbose XML data into a concise format that LLMs like Claude and Gemini can process with greater accuracy.Beyond development, the team shares their real-world applications of AI assistants, like John generating precise 3D printing step files based on scanned images and measurements.

The Portage Bay team sits down for a conversation about the tools and workflows that support their day-to-day work. Their discussion offers a behind-the-scenes look at how the devs approach problem-solving, collaboration, and efficiency as they kick off the year.

The team dives into the systems and tools they use regularly, like FM Perception, Base Elements, and 2Empower which support code navigation, debugging, and understanding complex data flows. While each developer brings their own preferences and opinions on which systems work best, these different approaches help the team collaborate to effectively and consistently deliver well-built, reliable software.

We had the pleasure of hearing Charles Delfs break down HAM (Headless Authorization Module) as he showed us how this system transforms the way you manage roles, permissions, and feature access in FileMaker. Inheritance, overrides, and data-driven rules can replace brittle security setups and open use cases like feature flags, SaaS plans, and time-based access.

We’re diving into the latest updates to FM Better Forms and taking a closer look at one of our own client’s web apps built on the platform. We share what’s new in Better Forms and how we’ve used it to build modern, real-world web apps that bridge the gap between low-code speed and polished web performance while combating challenges like caching.

Virtual lists are one of the most powerful and flexible techniques in FileMaker, giving developers the ability to generate custom reports, create dynamic pickers, and display complex data without storing it in fields.The team breaks down the essentials, like how global variables and JSON arrays feed into unstored calculations, how scripts can define headers and build arrays with SQL or looping, and why this approach is so effective for reporting and beyond.We also look at advanced use cases, from scaling large data sets and handling character limits to applying conditional formatting, integrating with WebViewer for interactive tables, and exploring Kevin Frank’s MD5-hashed window method for running multiple virtual lists at once.