Summer is here — and so is our favourite tradition. The 13.37% hacker discount is back. But this year we are going further: you can stack a second 13.37%, and we are being radically transparent about how — and whether — AI ever touches your data.

1. The summer hacker discount: 13.37% off everything

Through the summer we are taking 13.37% off all of our services — penetration tests, security audits, red teaming, code reviews, the lot. The rules are simple:

• Send us a binding order between 1 July and 31 August 2026, and
• we start the work before 31 August 2026.

That is it. Lock in your summer slot, get your assessment done before the autumn budget crunch — and pay 13.37% less for it. In autumn and winter we are usually swamped and hard to book, so summer is genuinely the smart time to schedule.

2. Go AI-augmented and stack another 13.37% (≈ 25% off in total)

Here is the new part. If you are happy for us to use publicly available frontier AI — Claude Opus 4.8 or GLM 5.2 — as part of your engagement, we will take off a second 13.37%. Stacked on top of the summer discount, that is roughly 25% off (13.37%, and then another 13.37% on what is left).

Why do we reward it? Because modern AI genuinely speeds up parts of our workflow — recon analysis, triage, reasoning about exploit paths — which lets us move faster and dig deeper in the same amount of time. When you opt in, we share that efficiency straight back with you as a discount.

3. “But you are sending my data to an AI?!” — No. Meet Anonproxy.

We know exactly what you are thinking, because we thought it first: handing pentest data to a third-party LLM sounds like a nightmare. So we do not.

Our Nethemba colleague John built Anonproxy — a reversible anonymization layer for sending data to LLMs, engineered to be far more reliable than naive match-and-replace.

Anonproxy sits between our tools (Claude Code, the OpenAI SDK, Burp Suite) and the LLM API. Before anything leaves the machine, it strips out IP addresses, hostnames, credentials, hashes, tokens, organisation names and PII — and then restores the real values in the reply. The model reasons over realistic surrogates; we read the real data. You never have to take “trust us” on faith — this is how we show you that your sensitive data stays yours.

It is open source: https://github.com/jfreemansh/Anonproxy

That is what transparency means at Nethemba: we can tell you precisely what does, and does not, leave your environment.

4. While you are here: free hacker videos

Sharpening your own skills? Our colleague Ian Budd keeps publishing free educational hacking videos on YouTube (the sabretoothatNethemba channel), spanning beginner to advanced. Great watching for your blue and red teamers alike.

Ready to save ~25%?

Get in touch via our contact form and mention the summer 13.37% discount (and whether you would like the AI-augmented stack on top). Just remember the two conditions: order by 31 August 2026, and we start before then. See you in the logs. 🛡️