Getting Started¶
In plain English
CortexBuild reads your existing codebase, writes down the unwritten rules your team already follows, and hands those rules to your AI coding tools so they stop breaking your conventions. This section gets you from zero to a repo that's been onboarded.
CortexBuild is a command-line tool you run against your own repository, on your machine, under your own provider keys. Nothing is sent anywhere you don't control.
The 3-step path¶
Follow these in order. Each step links to a focused page.
| Step | What you do | Page |
|---|---|---|
| 1. Install | Get the CLI running. | Install |
| 2. Onboard | Analyse your repo into .cortexbuild/ artefacts. |
Onboard Your Repo |
| 3. Render | Write the rules out to Copilot, Claude, Codex & Cursor. | Onboard Your Repo → step 4 |
New here? Start with the concepts¶
If you want the mental model before touching a terminal, read What is CortexBuild? first — it explains the problem, who it's for, and the rules-as-source-of-truth idea in one screen.
Executable hooks¶
Beyond the rendered files above, cortexbuild render also wires up
executable hooks for hosts that support them — session-start build
checks, turn-end telemetry, and a legacy/** edit guard the AI host runs
itself. See Executable Hooks for the full per-host
coverage table.
Then go deeper¶
Once your repo is onboarded, learn how the analyser works and how to roll the harness out to a team:
- How it works — what the analyser reads and writes, and how rendering works under the hood.
- Org Rollout — taking it to a whole team.