Overview

This section covers what OpsPilot is, the estates it is built for, who uses it, and what happens between your question and a command running on a server. Read it before you install anything. If you are reviewing OpsPilot for a security sign-off, start with Architecture.

The pages in this section

• What OpsPilot is — the definition, the principle the product is built around, and how it differs from an AI coding tool.
• Why OpsPilot — the capabilities that follow from reaching disconnected estates without giving a model a shell, with the mechanism behind each one and its limits.
• Who it is for — six roles, and what each one does with OpsPilot day to day.
• How it works — the eight-step loop from question to executed command, the two AI modes, and where each kind of data is held.
• Architecture — process separation, trust boundaries, data flow, and where configuration and credentials live. Written for an architect or a security reviewer.

See also

• Quickstart — first connection to first AI answer in about five minutes
• Execution boundary & risk tiers — the boundary the rest of the safety model rests on
• Glossary — definitions of the terms used across these pages