Why We Chose the MIT License for SafeClaw

Authensor Team · 2026-02-13

Why We Chose the MIT License for SafeClaw

Choosing an open source license is one of those decisions that seems simple until you sit down to actually make it. Every license embodies a philosophy, a set of trade-offs, and a bet about what matters most. For SafeClaw, we chose the MIT license. Here's why.

The Options We Considered

We seriously evaluated four licenses:

GPL v3 — The "copyleft" option. Requires derivative works to be open source. Ensures that improvements to SafeClaw always come back to the community. The trade-off: many companies have policies against using GPL-licensed dependencies, which would limit adoption. Apache 2.0 — A permissive license with patent protection. Popular with enterprise software. The trade-off: slightly more complex than MIT, with requirements around NOTICE files and patent grants that add overhead for contributors. Business Source License (BSL) — A time-delayed open source license used by companies like MariaDB and Sentry. Restricts commercial use until a specified date, then converts to a fully open license. The trade-off: not technically open source, which undermines the trust argument. MIT — The simplest permissive license. Do whatever you want with the code, just keep the copyright notice. No copyleft, no patent clauses, no restrictions.

Why MIT Won

Our decision came down to three factors:

1. Maximum Adoption

SafeClaw's value increases with adoption. Every user who deploys SafeClaw is one more developer team protecting their AI agents. Every additional user generates feedback, bug reports, and use cases that improve the product for everyone.

MIT is the most adoption-friendly license. No legal review required at most companies. No copyleft concerns. No patent implications. A developer can add SafeClaw to their project without consulting legal, which means more developers will add SafeClaw to their projects.

For a security tool, breadth of adoption is itself a security benefit. Widely-used security tools get more scrutiny, more testing, and more investment than niche ones.

2. Trust Through Simplicity

SafeClaw is a security tool that asks users to trust it with access to their development environment. Every barrier to trust works against us. A complex license creates doubt — "what are they trying to protect?" or "what restrictions am I agreeing to?"

MIT is 21 lines long. There's nothing hidden, nothing ambiguous, nothing that requires legal interpretation. The simplicity of the license mirrors the transparency we strive for in the software itself.

3. Our Business Doesn't Depend on the License

Some companies choose restrictive licenses because their business model depends on controlling the software. Our business model is built on the services and enterprise features we provide around SafeClaw, not on controlling who can use or modify the core product.

If a company takes SafeClaw's code and builds something with it, that's a success, not a threat. It means more AI agents are running with safety guardrails. That's the outcome we're optimizing for.

What We Give Up

The MIT license means competitors can fork SafeClaw, modify it, and sell it without contributing back. Cloud providers could offer SafeClaw-as-a-service without paying us anything. Companies could strip our name off and claim it as their own.

We're aware of these risks. We accept them because:

Most companies that fork don't maintain their forks long-term. The maintenance burden brings them back to the upstream project.
Our brand, our community, and our velocity are competitive advantages that can't be forked.
If someone builds a better version of SafeClaw, that's good for the mission. AI agent safety is not a zero-sum game.

The Philosophical Argument

We build safety tools. Our mission is to make AI agents safer for everyone. Restricting who can use our tools would directly contradict that mission. The MIT license is the most consistent expression of "we want everyone to be safe."

An AI agent running without guardrails is a risk, regardless of who's running it or whether they've paid us for a license. MIT ensures that there's no financial or legal barrier between any developer and the safety their agents need.

SafeClaw is on GitHub under the MIT license. Use it, fork it, modify it, sell it. Just make your agents safer. That's the deal, and we think it's a good one.

Read more about our approach to open source in our documentation.