Open source

We publish the guardrails.

Trust is earned before the sales call. We open-source the tools and guardrails — security, compliance, and reporting first — and keep commercial products and provenance internals private. Here's what's public.

Flagship · MIT

Local81

A lean, operator-readable deploy & runbook control plane: plan, deploy, audit, hold the line. Dry-run-first by default, with compliance and hardening reporting baked in.

Python 3.12CLIdeb / rpm Compliance scannersMulti-DB opsCI

Compliance & reporting, built in

Read-only operational checks mapped to recognized control themes, with inventory, findings, and advisory hardening plans you can hand to the next operator. Reporting that survives an audit.

AI

Security layer for self-hosted AI

We contribute to privacy-first, self-hosted AI projects like tinyhumansai/openhuman — our angle is the security, compliance, audit, and reporting layer that makes self-hosted AI safe to run in serious environments.

Coming soon

A standalone EDI/X12 validator core

A clean, self-hostable validator — free to run, easy to extend. The on-ramp to the full platform.

Watch the org →

Publish the tools. Keep the secrets.

Open guardrails build trust; private internals protect the moat. That balance is the whole strategy.