The Meeting That Started FlowForge

In July 2023, I sat down for a call with two representatives at Safe Software. I'd been an FME user for years, an FME Certified Professional, and honestly, a bit of an evangelist for the product. When colleagues asked how to solve spatial data problems, FME was usually my first recommendation.

The goal of that meeting was simple: renew my personal license at a reasonable price, like I'd done in the past.

The best price Safe would offer me was $7,300 per year.

I don't blame Safe for that. They're a business that can set whatever prices they want. But sitting there after the call ended, I realized something had fundamentally changed—not just at Safe, but in how I needed to approach spatial work going forward.

I made three decisions that day.

Decision One: FME Was No Longer an Option for Personal Work

At $7,300 annually, FME stopped making sense for the kind of work I do outside my day job. Side projects, consulting, experiments, building tools for the community—none of that pencils out when you're paying enterprise software prices as an individual.

This wasn't an emotional decision. It was math. And the math didn't work.

Decision Two: No More Evangelizing

This one was harder. I'd spent years recommending FME to colleagues, teaching people how to use it, advocating for it in organizations that were evaluating spatial ETL options. I genuinely believed it was the best tool for the job.

But I couldn't keep recommending a product that had priced out independent professionals and small teams. When someone asked me "what should I use for spatial ETL?" I needed a different answer. The problem was, I didn't have one yet.

Decision Three: Convert Everything to Code

If I wasn't going to use FME, I needed an alternative. The immediate solution was obvious: Python.

GDAL, GeoPandas, Shapely, Fiona, PyProj—the libraries exist. They're mature. They're free. I started converting my FME workbenches to Python scripts, one workflow at a time.

It worked. Mostly.

But if you've done this, you know the friction. Every workflow is a new script. There's no visual feedback. Debugging means print statements and log files. Sharing with colleagues who aren't coders means wrapping everything in a CLI or a notebook and hoping they can follow along.

I was trading licensing costs for development time. For some workflows, that tradeoff made sense. For others, I found myself missing the visual canvas, the drag-and-drop, the immediate feedback of seeing your data flow through transformations.

The Decision I Didn't Say Out Loud

There was actually a fourth thing running through my head after that call, but I didn't commit to it immediately. It felt too ambitious.

What if I built my own alternative?

Not a Python wrapper. Not a notebook framework. A real desktop application with a visual workflow builder, native performance, and no annual licensing fee that costs more than some people's car payments.

I'd been writing software for 30 years. I knew Rust. I knew the spatial ecosystem. I knew exactly what I wanted FME to be and wasn't.

So I started building.

Two Years Later

FlowForge exists now. It's a Rust-based spatial ETL tool with a visual workflow builder, DuckDB under the hood for analytical performance, and a Tauri desktop shell that doesn't eat your RAM like Electron would.

Let me be clear: FlowForge doesn't have the breadth of functionality that FME has. Not yet. FME has decades of development and hundreds of transformers. I'm one developer who started in 2023. But the core workflows—the things I actually needed to do every day—those work. And they work fast.

It's not trying to be FME. It's trying to be the tool I needed in July 2023—something powerful enough for serious spatial work, accessible enough that you don't need to be a programmer to use it, and priced for humans rather than enterprise procurement departments.

I still respect FME as a product. The engineering is solid, the format support is unmatched, and for organizations with the budget, it's a legitimate choice. But the spatial data community is bigger than enterprise customers. There are independent consultants, small shops, researchers, hobbyists, and professionals who need spatial ETL but can't justify $7,300 a year.

FlowForge is for them. FlowForge is for me.