Closed source developer tools are a contradiction.

You’re asking a developer — someone who reads code for a living, who questions every dependency, who opens the source of any library before they ship it — to trust a black box. That doesn’t work. It never worked. It just used to be the only option.

It’s not the only option anymore.

The Trust Problem

Developer tools sit deep in the stack. They touch your codebase, your terminal, your file system, sometimes your credentials. The closer a tool gets to the sensitive parts of your workflow, the more you need to know what it’s doing.

Closed source at that layer isn’t just philosophically uncomfortable. It’s a practical problem. You can’t audit it. You can’t verify the claims. You’re taking a vendor’s word for what happens to your data, your context, your prompts.

That’s a bad deal. I took that deal for years with various tools. I’m done taking it.

Open Source Isn’t a Business Decision

I keep reading that open sourcing is a “strategy” — a way to build distribution, reduce CAC, attract contributors. Maybe. But that’s not why I’m doing it.

I’m doing it because the tool I’m building lives on your machine, connects to your dev environment, and reads your code. You should be able to see exactly how that works. Not because I’m legally required to. Because it’s the right thing to do.

If the code is bad, you’ll see it and tell me. If something looks wrong, you can check. If a future version does something you don’t want, you can diff it. That’s the relationship that makes sense between a developer and the tools they use every day.

The Irony of AI + Open Source

There’s something uncomfortable about building an AI coding tool on top of open source infrastructure while shipping it as a closed product. I’ve written about AI strip-mining open source — consuming without contributing. I’m not interested in being part of that pattern.

Open sourcing isn’t a complete answer to the sustainability problem in OSS. But it’s at least not making it worse.

Local-First and Open Source Go Together

Privacy-first, local-first tools have a credibility problem if they’re closed. “We don’t send your data anywhere” is a claim. Open source is evidence.

These things belong together. If the whole point is that you own your environment, your data, your context — then you should also be able to own the code that runs it.

What This Means

Everything goes up. The daemon, the mobile app, the protocol. No “open core” with a paid moat around the good bits. No community edition that’s deliberately crippled.

If you want to run it yourself, you can. If you want to change it, you can. If you want to audit it before you trust it, you should.

That’s the deal.