ArmorIQ Tools
Guardrails and a flight recorder for AI coding assistants. One command to install, then see and control everything Claude Code or Codex does on your machine.
AI coding assistants do not just suggest code. They read your files, edit them, run shell commands, and call out to the internet, often faster than you can follow. ArmorIQ Tools gives you two things:
- A record. Every action your assistant took, in order, with what it touched.
- A say. Rules that stop or pause the actions you do not want it taking on its own.
Install it with one command. You choose what gets stopped, and you can start by just watching.
curl -fsSL https://armoriq.ai/install_armorclaude.sh | bashThat is for Claude Code. For Codex, use install_armorcodex.sh. Then open
tools.armoriq.ai to watch runs arrive.
Start here
Quickstart
Install, approve your machine, and get your first recorded run. About five minutes.
How it works
What actually happens to a tool call, what a session records, and what a week tells you.
Tour of the console
Every screen at tools.armoriq.ai and what each one answers.
Writing rules
Policy Studio: allow, hold, or block a tool, and lock a folder.
Is this for me?
It fits if you use Claude Code or OpenAI Codex and any of these is true:
- You want to let the assistant work on its own, but not blindly.
- Something changed and you need to know what the assistant did and when.
- You have a folder, a credential file, or a production config you never want touched.
- You are paying for tokens and want to see where they went, by session and by machine.
It is built around a single workspace per account today. There is no team management and no invites, though billing does exist in Settings → Billing. See Limits for the full list of what is not built, so you can decide with the real picture.
Everything else sits under Reference in the sidebar: signing in, API keys, enforcement, telemetry, architecture, and limits.