Services

Every engagement is scoped with you.

There are no packages, and you don’t need to know what to ask for. We talk through what you’ve built and what worries you, and we judge where an attacker would get the most out of it. The test is built around that. The price is agreed before testing starts and does not change.

01 / Where to start

Start from where you are.

Not sure what you need? Most teams aren't. These are the situations we hear most often, and where we would usually begin.

  1. 01

    You've never had a test and don't know where you stand

    We start with your highest-risk surface: usually where people sign in, where money or personal data moves, and whatever is open to the internet. You come away knowing where you stand and what to fix first.

  2. 02

    You're about to launch, or a customer is asking for a pentest

    We test what's going live, the web app and the APIs behind it, and give you a report you can share with the customer who asked.

  3. 03

    You've added an AI feature

    We test the feature and everything it can reach: what it can be talked into, which tools and data it can touch, and what it could leak.

  4. 04

    An audit or regulation asks for testing

    DORA, NIS2, ISO 27001, SOC 2. We scope the test to what the requirement asks for, and write the report for the auditor as well as your engineers.

02 / What we test

Three areas, in whatever mix your app needs.

Most apps touch more than one. How much of each, how deep, and whether it is one test or a recurring one is what the first call is for.

  1. 01

    Web applications and APIs

    Your web app and the APIs behind it, tested as each of your user roles: authentication, authorization, business logic, data exposure and tenant isolation.

  2. 02

    AI and LLM features

    Chatbots, copilots and agents that call tools: prompt injection, data leaking through the model, tool abuse and excessive agency.

  3. 03

    Cloud trust

    How your app reaches its cloud: roles, keys, service accounts and the paths between them, on AWS, Azure or Google Cloud.

03 / Code access

Whitebox or blackbox.

Whitebox testing requires read access to your source code. We settle which one fits on the first call.

  1. 01

    Whitebox, with source code access

    We read how your code handles data and permissions, then prove what we find against the running app. It goes deeper in the same time, down to flaws a blackbox test cannot reach. This is what we recommend.

  2. 02

    Blackbox, without code access

    We test the running app from the outside, the way an attacker would. Useful when code cannot be shared, but it sees less in the same time.

04 / Included

The same terms for every engagement.

However we scope it, it runs the same way and ends with the same deliverables.

01

An intro call

We work out together what is most at risk, what to test first, how deep, with what code access and when. Then we quote for exactly that.

02

Written authorization

Signed before any testing, naming targets, dates and contacts.

03

One lead, start to finish

The same researcher scopes, leads, reports and debriefs. Agents work where we send them, and nothing they turn up reaches your report until a researcher has checked it.

04

Same-day critical findings

Anything critical reaches you the day we find it.

05

Report with proof and fixes

Ranked findings, steps to reproduce and a fix for each.

06

Debrief and retest

A call with your engineers, then a retest once the fixes are in.

05 / Questions

Questions teams ask us.


  1. 01

    We already run an AI security review on our code. Why do we need you?

    Keep running it: it catches problems early. It reads your code and suggests what might be wrong. We test your app as it actually runs, including the frameworks and libraries underneath it, and a researcher verifies every finding before it reaches your report, with its impact and a fix.

  2. 02

    We don’t know what to test. Can we still talk to you?

    Yes. Most teams we talk to don’t. The intro call is free and takes thirty minutes: tell us what you’ve built, and we work out where an attacker would get the most out of it and what to test first.

  3. 03

    Do you need our source code?

    We recommend it. With read access to your code we go deeper in the same time. If it can’t be shared, we test the running app from the outside.

  4. 04

    Does our code go to an AI?

    Yes, to open-source models served by EU providers in EU datacenters. Never to a closed model behind an AI vendor’s API, and your code and data are never used to train a model.

  5. 05

    Is anything tested before we sign?

    No. Nothing is tested until you sign a written authorization that names the targets, the dates and who to call.

Quote

Get a fixed quote.

The call and the quote are free and do not commit you to anything.

Book a call with us