What “Kali GPT” actually is

Short answer: there is no official “Kali GPT” from Kali Linux or OffSec. The name is used by several unrelated things: a paid product at kali-gpt.com, a custom GPT on the ChatGPT store, an open-source CLI on GitHub, and shell-gpt (which is in Kali’s repos). What they have in common is an LLM wired into your terminal. You can build the same thing, privately, by pointing an OpenAI-compatible CLI at a cybersecurity model on Obscurify.ai — here’s the setup.

If you search “Kali GPT” expecting an official, more-permissive AI baked into Kali Linux, you won’t find one. OffSec and OpenAI have announced nothing of the sort. As one write-up tracing the hype put it, much of it traces back to ordinary GPT wrappers. Here’s what the name actually points to.

The things called “Kali GPT”

1. kali-gpt.com — a paid product. A Shopify store selling “Kali GPT” as a one-time purchase (around €79.90), described as a GPT-4-powered copilot for Kali tools (Nmap, Metasploit, Burp, Aircrack-ng, Hydra, Hashcat). It is not affiliated with Kali Linux or OffSec, and its testimonials read like stock marketing. Treat it as a third-party assistant, not official tooling.

2. A custom GPT on ChatGPT. KaliGPT on the ChatGPT store is a custom GPT — a system prompt on top of OpenAI’s models, which anyone can make. It inherits ChatGPT’s refusals, so it still declines plenty of legitimate security questions.

3. An open-source CLI. SudoHopeX/KaliGPT (“HackerX”) is an actively developed agentic CLI for Linux that calls Gemini, OpenAI, Ollama or any provider through LiteLLM, with tool calling and Termux support. This one is genuinely useful and model-agnostic — you choose the backend.

4. shell-gpt (sgpt) — the real Kali package. shell-gpt is a general LLM command-line tool that actually ships in Kali’s repositories (sudo apt install shell-gpt). It isn’t security-specific, but it’s what many “GPT in your Kali terminal” guides really use, and it takes any OpenAI-compatible endpoint.

5. Credulous coverage. Some articles (for example on LinuxSecurity) describe “Kali GPT” as a trained GPT-4 assistant, which blurs the line further. There’s no single official tool behind the name.

What people actually want from it

Strip away the branding and the want is simple: an LLM in your Kali terminal that helps with recon, tool flags, reading output and writing PoCs for authorized testing — without refusing legitimate security questions. The pieces that deliver that are (a) a terminal CLI that speaks to an OpenAI-compatible endpoint, and (b) a model that engages with authorized offensive work. The CLIs above all do (a). The gap is usually (b): a wrapper over a mainstream model still hedges.

How to get a private equivalent

Point one of the terminal CLIs at a cybersecurity model on Obscurify.ai. Both common options take an OpenAI-compatible endpoint:

The step-by-step is in the Kali GPT setup tutorial. The result is the thing the name promises — an AI copilot in your Kali terminal — but on a model tuned for security work, on an account that is just a random 16-digit number, with nothing stored or trained on.

Why a cybersecurity model, not a wrapped ChatGPT

A custom GPT or a wrapper over a mainstream model keeps that model’s refusals, so it balks at authorized recon scripting, exploit PoCs or malware analysis. Obscurify.ai’s cybersecurity-tuned models — the OrcaCyber family and CyberHeabsy — are built for that work and run on any plan, including Pay As You Go with no monthly fee. See the cybersecurity LLM page for the list, the uncensored coding AI page for day-to-day coding, and our open refusal test for how the broader uncensored models behave.

A note on scope

This is for authorized work only: pentests you have permission to run, CTFs, bug bounty within scope, reverse engineering, malware analysis, detection engineering, OSINT and education. Obscurify.ai’s terms prohibit writing or spreading malware and breaking into systems you aren’t authorized to access. An AI that answers legitimate security questions is not a licence to do illegal work.

Start

Get an account (a random 16-digit number), fund it with Monero, Bitcoin, Lightning or a card, and follow the Kali GPT setup tutorial. New accounts get $1 of starter credit, with the Pro models (uncensored ones included) for their first 7 days, while this month’s promo budget lasts.

Sources

  1. kali-gpt.com (paid product): https://kali-gpt.com/
  2. KaliGPT custom GPT: https://chatgpt.com/g/g-xouSQobsE-kaligpt
  3. SudoHopeX/KaliGPT (open-source CLI): https://github.com/SudoHopeX/KaliGPT
  4. shell-gpt in Kali: https://www.kali.org/tools/shell-gpt/
  5. “What the fuck is kali-gpt?” (origin investigation): https://hack4living.com/1337/06/08/what-the-fuck-is-kali-gpt/

Obscurify.ai facts: pricing, privacy policy, terms, cybersecurity models.

Common questions

Is Kali GPT an official Kali Linux tool?

No. Neither Kali Linux nor OffSec publishes a tool called “Kali GPT”. The name is used by several unrelated third-party products and projects. The only AI CLI that actually ships in Kali’s repos is shell-gpt.

Is the kali-gpt.com product worth buying?

It’s a third-party paid assistant (around €79.90), not official and not affiliated with Kali or OffSec. You can get the same “AI in your Kali terminal” result for free with shell-gpt or the open-source KaliGPT CLI, pointed at the model of your choice.

How do I use a model that doesn’t refuse in Kali?

Install a terminal CLI that takes an OpenAI-compatible endpoint (shell-gpt or SudoHopeX’s KaliGPT), set the base URL to https://obscurify.ai/v1, and pick a cybersecurity model. The tutorial walks through it.

Does it keep or train on my prompts?

No. On Obscurify.ai, prompts and replies aren’t stored on the server, and we don’t train on it, and models with the ZDR tag only go to zero-data-retention routes, whose providers’ published policies say they don’t train on or keep the data (published policies, not signed agreements). CyberHeabsy has the tag; the OrcaCyber models don’t, so their provider’s own policy applies. The account is a random 16-digit number with no email.

Which model should I use?

For security work, a cyber-tuned model like OrcaCyber K3.1 Pro or CyberHeabsy. For writing tools and PoCs, a coder model. They’re on the same account and API.

Last updated 2026-11-08. Plan prices and model lists on this page are read live from the site.

Try it now: guest chat needs no account, no email and no cookies. When you want more models, an account is just a random number.

Kali GPT setup tutorial See pricing
Kali GPT setup tutorial