Why Cascadia PLM Is Open Source
My Frustration With the Enterprise Software Industry
I've been working in PLM for six years now. Feels like longer. Before that I spent about 8 years in various mechanical engineering and product management related roles, so I came into PLM as many of the best PLM solution architects do, with a strong personal understanding of the engineering user's perspective.
And then I was put in charge of my company's PLM, CAD, and eventually entire software stack (ERP, CRM, manufacturing software, the whole 9 yards). So I saw firsthand what enterprise software costs a business. Then I moved on, and saw that same business from the consulting side, and then from the independent contracting side, and then finally from the large enterprise white-label side.
And there were pieces of that puzzle, big pieces, that could only be explained by a predatory market situation. A few old monsters (Siemens, Dassault, PTC) control the Tier 1 PLM industry. Their offerings are the only options at the highest levels (that's a topic for another blog), but those highest levels are there to serve large customers with specific needs for functionality only offered by those top tier offerings, and where the subscriptions and consulting fees easily extend into the millions of dollars a year just to minimally maintain and manage "just good enough" deployments into those environments.
Everyone else is forced to overpay for functionality they don't need, or take a step down into Tier 2 or Tier 3 PLM offerings. But even that's a trap. Those Tier 2 offerings can easily still run a company six to near seven figures in subscription costs alone. And no company gets away from the need to customize the solutions offered to them in some way. Those customizations inevitably have to be done by very specialized consulting shops, and those engagements are often ongoing, sometimes off and on, for the life of the application, and those engagements can easily cost more than the subscription fees for the software itself!
Then there's the burden of having an in-house developer or highly technical "admin" for each of your systems. That additional cost can again run more than the subscription fees alone, and a lot of companies try hard to avoid staffing that role, only to find later on that the software they've bought and customized is basically useless without someone constantly propping it up internally.
What Cascadia Aims to Do
With the advent of cheap, relatively high quality software development thanks to the advances in LLMs beginning around November 2025, and tools like Claude Code, Codex, and Cursor, I decided it was time to try to democratize PLM.
Inspired by projects like carbon.ms, I'm releasing Cascadia PLM to the world. An open source PLM platform, Cascadia PLM is free to self-host with unlimited users and all features included. There's no open-core gating, no premium tier that holds back critical functionality. The full platform is available on GitHub under AGPL-3.0. Pull it down, run the build on your own server, and get to work.
There are some (entirely optional) pieces of it that you will need to bring your own API keys to (for AI-assisted design pipelines, or automated CAD-generation), but you can freely ignore those looking-forward features, and simply use it as a traditional PLM.
Cascadia's Focus
While I think that Cascadia is suitable for any PLM workflows, I built it to address what I've seen as a gap in the current market specifically: Engineer-to-order (ETO), or high mix, low volume (HMLV) manufacturing companies, specifically in the small-medium business (SMB) segment.
Cascadia brings a focus on speed and flexibility of engineering development (through a git-inspired change management workflow), and integrated engineering-to-manufacturing workflows (EBOM to MBOM + work instructions and work orders built in, for companies that don't want or need a complete MES platform quite yet). More posts will be coming about the specific design choices, so check back in here from time to time.
And if any of that interests you, check the public repo on Github, follow along, open an issue if you find something broken (and we're early, still, things will be broken), and reach out via email or linkedin. I welcome all feedback!