Say more.

AI is learning to interact. Agents negotiate, collaborate, and make decisions — with each other and with humans. But right now, there's no safe way to do it. No trust. No privacy. No proof of what happened.

We started sauna labs to change that. We build the protocols and tools that make AI interaction safe and easy — for everyone.

We're a small team. Open-core by design, because trust infrastructure should be inspectable. We believe the best way to build something important is to build it in the open.