Case study · P4 One · Version control for creative professionals
P4 One is Perforce’s version control app for artists and designers. I introduced Experiment, a way to try ideas and safely go back. It shipped in Unreal Engine. I then designed and built it for Blender, where it is planned to ship in 2027.
Version control tracks changes to files, and most artists don’t use it. We didn’t want to build one more place to dump files. We wanted to protect their work inside their own workflow.
Artists find version control confusing, and leaving their tool to save means they often skip it.
“Can I safely get back to where I was?”
Introduced Experiment, then designed and built it for Blender in Figma, VS Code and Claude, working with our researcher, PM and engineers.
Experiment shipped in Unreal Engine, free on Fab. The Blender plugin got positive feedback from 14–16 users and is planned to ship in 2027.
A plugin inside Blender, not another app to open.
Press N to show it, press N to hide it.
Not a mini desktop app.
First in Unreal Engine, free on Fab, where a couple of people left feedback. The Blender plugin got positive feedback from 14–16 users, gathered in research sessions our researcher ran and in conversations our PM had with users. It is planned to ship in 2027.
The idea grew into a vision: one panel across the creative tools. Honorable mention, Perforce Jam, May 2026. See the Jam project
I’d study artists’ workflows before designing anything for the desktop app, which would have saved us a lot of time. And I’d plan the app and the plugins as one product from the start, because testing separately in Unreal and Blender left the pieces fragmented.
The full story, in short
Artists rarely use version control. It’s confusing, and saving means leaving their tool.
Polishing versioning didn’t help. The real need was to try ideas safely, inside their own tool.
Experiment shipped in Unreal Engine, free on Fab. The Blender plugin I designed and built is planned for 2027.
After we acquired Snowtrack, a desktop app for creatives, we were asked to polish its versioning. Better metadata still left it as just another place to store files.
We spoke with 6 artists and designers. Every experiment became another file, and saving meant leaving their tool, so they often didn’t.
The folder on the right is an illustrative week, not real data.
One question kept coming back. It changed the project: from storing versions to protecting experiments.
My three design pillars
One-click core actions.
Advanced actions only when needed.
Always-visible feedback and confirmations.
Experiment lets you try anything and safely go back, inside Blender. Here is the real panel.
Engineers built a quick first plugin with a modal. After research, I proposed a panel in Blender’s N menu: press N to show it, press N to hide it.
Schematic of the two options, not the real screens.
I wrote the plugin’s scope. I wanted to do more, and we agreed on what it does best.
The live, day-to-day work
Advanced tasks. The plugin opens it.
Creating says it won’t touch your main work. Deleting says exactly what is lost, and that it can’t be undone.
I introduced the Experiment concept. Engineers built a simple first version in Unreal Engine and it shipped free on Fab: the fastest way to learn whether the concept worked.
Try it yourself A small playable version of Experiment, built with AI. Watch the version soup get fixed, then switch between ideas.
every artist knows version soup
An illustrative week of files and a toy car, not the real plugin or real data. The experiment names come from my sample project.
I designed the flows in Figma, then built the plugin live in Blender with Claude and VS Code. AI confidently invented things and cleanup cost time, but it proved feasibility fast.
VS Code + Claude 14–16 users, positive feedback
First in Unreal Engine, free on Fab. The Blender plugin I designed and built got positive feedback from 14–16 users and is planned for 2027.
The idea grew into a vision: one panel across the creative tools. Honorable mention, Perforce Jam, May 2026. See the Jam project
I’d study artists’ workflows before designing anything for the desktop app, which would have saved us a lot of time. And I’d plan the app and the plugins as one product from the start, because testing separately in Unreal and Blender left the pieces fragmented.