Rant AI CLI

An AI cybersecurity agent for your terminal, by CyberRant. Ask in plain language; Rant AI runs real security tools on your machine and answers with the result — nothing more.

cyberrant@rantai:~# can you check my network IP and storage?
  ✓ Done  1.2s

  Network
  Public IP: 203.0.113.7
  Local IP: 192.168.1.20

  Storage
  Used: 320.0 GB (67.3%)
  Free: 155.4 GB

Requirements

• Node.js 18 or newer
• Python 3.10 or newer, on your PATH
• Internet access during install (Python dependencies are downloaded once)

Install

npm install -g @cyberrant-rantai/rantai

The install step creates a private Python environment inside the package and installs its dependencies. This takes a few minutes the first time.

Use

cd your-project
rantai

Sign in with your CyberRant account on first launch. Rant AI works in the folder you start it from.

Requests that need AI (planning, answers, AI scans) run on CyberRant's servers and are limited to 25 per day per account; the count resets at midnight UTC. One prompt is one request, however many steps it takes. Built-in actions don't use one: IP and disk checks, the current folder, project summaries, scans without --ai, and plain commands like git status.

Things to try:

• what is my public ip?
• how much disk space do I have?
• scan this project for vulnerabilities and exposed secrets
• check the SSL certificate and security headers of example.com
• fix the security vulnerabilities in this project — every file change is shown as a diff for you to accept or reject, and can be undone

Anything that runs a command or changes your system asks for permission first:

  Rant AI wants to run:
    • Listing the files in src   dir /b src
  Allow? (Y/n · d = details)

Commands

| Command | What it does | |---|---| | /help | All commands and example prompts | | /verbose | Toggle full internal output: plans, tool cards, raw tool output | | /trace | Show what ran on the last request, with timings | | /trace full | Same, with complete tool output and logs | | scan / audit / recon | Quick system scan / full audit / recon of a target | | report / dashboard | Save an HTML report / open the interactive dashboard for the last assessment | | history, clear, logout, exit | Session basics |

CI and monitoring

Run the security scanners without the interactive UI:

rantai --ci . --fail-on high --sarif results.sarif   # exit 1 if issues ≥ high
rantai --monitor . --once                            # one scan cycle, for cron

--ci exit codes: 0 clean, 1 findings at or above the threshold, 2 error. More options: --json <file>, --ai (AI audit — needs sign-in, uses one daily AI request per run), --no-cve, --no-secrets, --quiet.

Settings

| Environment variable | Effect | |---|---| | RANTAI_VERBOSE=1 | Start in verbose mode | | RANTAI_ASCII=1 / RANTAI_UNICODE=1 | Force plain or Unicode symbols if they render wrongly | | RANTAI_NO_OPEN=1 | Never open dashboards in the browser automatically |

Privacy and responsible use

• Public-IP lookups call api.ipify.org.
• Only scan systems you own or are authorized to test.

Links

• Website: https://cyberrant.org
• Issues: [email protected]