Nova: On-demand agentic pentesting

If someone went after our app, how far would they actually get?

Nova answers it, in hours. Point Nova at an application. A fleet of AI hacker agents works through it the way a skilled pentester would, step by step, until it can show real impact. Every finding comes with the evidence.

Trusted by companies globally

Questions Nova answers

Can one customer see another customer's data?

Nova checks, the way an attacker would. It signs up like a real customer and tests whether the line between accounts holds. If it doesn't, you see exactly how, request by request.

We launch on Friday. Is it ready?

Run Nova on staging first. Point it at the staging app before launch and you get a full pentest in days, not a slot in next quarter's calendar.

Can we get a pentest this week, not next quarter?

Start today. There's no SOW to negotiate and no team to schedule. Most tests finish within 24 to 48 hours, and you follow the results as they come in.

Case study

"Hadrian enables us to pinpoint the real security issues we should be working on."

Hans Quivooij, CISO

Damen Shipyards Group

Read case study
The findings are described in real detail, with enough reproduction information and context that we can go straight into remediation without having to guess at intent.

A manual pentest, and Nova.

manual Pentest
Hadrian Nova
scheduling
Weeks of coordination
On demand, start today
Scoping
Negotiation with vendor
You define it, you control it
Turnaround
Weeks to months
Hours
Consistency
Depends on the tester
Same quality every time
Coverage
Same playbooks each time
Evolves with attacker tradecraft
Validation
Varies by engagement
Human-reviewed, zero false positives
You get
A static PDF at the end
Live results, plus a compliance-ready report
Proof looks like
Varies by tester
Every request and response, attached
Price
Scoped per engagement
From €1,250 per test excluding VAT, with bundles

Launch a pentest in a few clicks.

Click through it in the product: set the scope, start Nova, and open the findings it proves, with the evidence attached.

Frequently Asked Questions

Learn how Hadrian helps security teams find, verify, and fix what attackers actually exploit.

What is a Nova pentest?

Nova is an agentic penetration test powered by AI. It performs the same techniques a skilled human pentester would, including reconnaissance, exploitation and lateral movement, but faster and with full transparency into every step taken.

How long does a test take?

Most tests complete within 24 to 48 hours depending on scope. You receive access to the Hadrian platform to see your results and get a compliance-ready pdf report.

What's included in the report?

Every finding includes proof of exploitability, reproduction steps, risk context and remediation guidance. Results map to compliance frameworks including SOC 2, ISO 27001 and NIS2.

How is Nova different from a traditional pentest?

Traditional pentests require weeks of scheduling, depend on individual tester skill and deliver a static PDF at the end. Nova launches on demand, tests consistently every time and gives you full visibility into the AI's reasoning, tests performed and evidence collected.

How much does a Nova test cost?

Each test starts at €1,250. Bundles are available for teams planning multiple tests. Reach out for a custom quote based on your scope and frequency.

See pricing page for more information.

Can I use Nova alongside Atlas?

Yes. Atlas provides continuous external exposure management, monitoring your attack surface and validating exposures around the clock. Nova adds depth with on-demand penetration tests scoped to specific environments or compliance requirements. Together they give you both breadth and depth.

Who is Hadrian?

Hadrian is an offensive security platform used by security teams at 300+ enterprises worldwide. We were named a Representative Vendor in the 2026 Gartner Market Guide for Adversarial Exposure Validation and a Leader 3 years in a row by GigaOM.

Point Nova at your app.

Find out how far an attacker would get, before one tries.