Intelligence, illuminated.

Security for AI agents and autonomous systems. See what they do, control what they can, and trust the result.

Request early access

The problem

AI agents now read, decide, and act with real credentials. Most teams can't see what they're doing, or stop them.


How it works

Put a control layer between your agents and the tools they use.

Any agent ANY FRAMEWORK Aluz CONTROL LAYER Identity · Policy Approval · Audit Tools and data APIS · FILES · MCP ALLOW · ASK · DENY Any agent ANY FRAMEWORK Aluz CONTROL LAYER Identity · Policy · Approval · Audit ALLOW · ASK · DENY Tools and data APIS · FILES · MCP
Illustrative architecture. Aluz is in development.

The idea

The more capable intelligence becomes, the more it needs light.

Power without visibility is a risk. We are building for the opposite: systems you can inspect, question and rely on, from the first day.

Aluz comes from the Portuguese a luz: the light.


In development

Be first to see it.

We're building Aluz now. Join the list and we'll write when there is something real to show, and not before.

Built for developers building AI agents and autonomous workflows.