PRAESIDIX · AI GOVERNANCE · COMING SOON

Your AI agents, with permission, with limits and with a record of everything.

Praesidium: the garrison that sits in front of your MCP servers, controls who gets through and keeps a record of everything.

LET'S DESIGN IT TOGETHER In the lab

IN SHORT

What is Praesidix?

Praesidix is a governance gateway for AI agents that use tools over MCP. It sits in front of your MCP servers and decides who can use which tool, with what identity and within what limits, and keeps an audit log that can’t be tampered with. It is in the lab: we are looking for organisations to design it with.

THE PROBLEM

What problem does it solve?

Your agents already touch real systems. Every tool they use reads data or carries out actions, and each MCP server handles permissions its own way, or not at all.
A tool can change without warning. If its description changes, what the agent does with it changes too.
You will be held to account. Who did what, with which permission and when, in a record you can trust.

GOALS

What will it do?

Goal	What it means
A single entry point	Registration, discovery and routing for all your MCP servers.
Identity without shortcuts	It validates every agent’s and user’s identity with your provider (OIDC) without forwarding their credentials.
Per-tool permissions	Which role can use which tool, and within what usage limits.
Pinned tools	Any change to a tool’s definition is detected, compared and held for approval.
Tamper-proof audit	A log designed for what ENS, DORA and the AI Act require.
Real isolation	Strict separation between customers, and operation on air-gapped networks.

Three security profiles: local, enterprise and high security. What it isn’t: an identity provider, a service mesh, a generic API gateway or a language-model proxy.

WHO IT'S FOR

Who is it for?

Organisations with AI agents in production or about to have them, especially in regulated sectors: banking and insurance, infrastructure and high-security environments.

STATUS

What stage is it at?

In the lab. We have finished the market and regulatory study; the next step is a prototype. If your organisation is about to put agents to work on real systems, we want to design it with you.

FAQ

What people usually ask us about Praesidix

What is MCP?

The Model Context Protocol, an open standard AI assistants and agents use to work with external tools: querying a database, opening a ticket, reading a document.

What is a tool «rug pull»?

When a tool changes its description or behaviour after you have approved it. Praesidix will pin every definition and ask you to approve any change.

Is it available?

Not yet. It is in the lab and we are looking for organisations to design it with.

Does it replace my identity provider or my API gateway?

No. It relies on your identity provider and lives alongside your gateway; its job is to govern what the agents do.

PRAESIDIX

Are you about to put agents to work on real systems?

Let’s design together how to govern them.

LET'S DESIGN IT TOGETHER