Skip to main content
Hashnode
MEHO - From the labMEHO - From the lab

About

About MEHO

MEHO is an open-source governance layer for AI agents acting on infrastructure. It sits between the agent and the systems it operates — VMware Cloud Foundation, vSphere, NSX, Kubernetes, Vault, Keycloak and others — and enforces three things before anything happens: a policy check on every operation, credentials the agent never holds, and an audit record that is written before an operation counts as done. Risky operations wait for a human. The most destructive ones wait for a human every time.

The name stands for Machine Enhanced Human Operator, which is also the idea: the human stays the operator, the machine makes them faster, and the system makes sure the machine cannot do more than it was allowed to.

About this blog

MEHO was built in evoila's own lab, where AI agents have been doing operational work on real infrastructure under supervision since spring 2026. This blog is written from that lab. We publish what we build, what breaks, what we got wrong and what we changed, with release notes read as decisions rather than feature lists. Where we quote numbers or incidents, they come from our own records or from sources we link.

We do not publish product pitches here. If you want to talk about using MEHO, the contact is below.

About evoila

evoila is a European IT consultancy and Broadcom Pinnacle partner focused on cloud infrastructure, platform engineering and automation, with teams in Germany, Bosnia and Herzegovina and several other countries. MEHO is developed and maintained by evoila and released under the Apache License 2.0.

Links