Pricing

Fixed scope. Fixed price.

Every price includes the report, the attestation letter addressed to a named third party, and one retest of remediated findings. VAT is added at checkout. Annual prepay is ten months for twelve. No per-domain, per-host, per-application or per-repository limits — the scope is what you own, not a seat count.

Baseline

Black box

External perimeter, no credentials. Every domain and host you own.
€2,400
or €445/month · €4,450/year prepaid
  • OWASP Top 10 + exposed services
  • TLS, DNS, SPF/DKIM/DMARC
  • Exposed admin panels, leaked credentials
  • Report + attestation + 1 retest
Choose Black box
Attested

White box

Source, CI/CD and cloud configuration. All repositories and cloud accounts.
€11,500
or €1,950/month · €19,500/year prepaid
  • Everything in Grey box
  • Source review across your repositories
  • Secrets and CI/CD review
  • Cloud configuration review
  • NIS2 / DORA / ISO control mapping
Choose White box
Comparison

What changes between scopes.

Depth

Black box tests what an outsider sees. Grey box adds the authenticated surface where most real weaknesses live. White box reads the code and configuration where logic and secrets hide.

What your auditor gets

All three produce the same report format and an attestation letter; white box adds a control-mapping appendix tying each finding to a clause, which is what NIS2 and ISO files ask for.

What none of them include

Denial of service, social engineering, physical access, third-party SaaS vendors, and remediation implementation. Each can be added by written agreement.

AI & LLM

AI systems assessment.

For products with a chatbot, an agent that calls tools, RAG over internal documents, or an MCP server. Same report, same attestation, different test plan.

Most chosen

AI & LLM assessment

Your AI features end to end, every attack path tested through the real pipeline.
€4,900 one-off
or €795/month · €7,950/year prepaid
  • Prompt injection: direct, indirect, cross-modal
  • Agent and tool permission abuse (excessive agency)
  • RAG and vector isolation between tenants
  • MCP / plugin supply chain and tool-description tampering
  • Unbounded consumption and cost-abuse testing
  • Improper output handling into DB, browser, shell, API
  • Report mapped to OWASP LLM Top 10 2026 + Agentic/MCP Top 10
  • Attestation letter + one retest
Scope an AI assessment

What you get that a scanner cannot

Reproducible prompts, captured model responses, the exact tool call that fired, and the fix at the layer where it belongs — application, not prompt. Every finding ships with evidence your engineer can replay.

Frameworks referenced: OWASP LLM Top 10 2026, OWASP Agentic & MCP Top 10, MITRE ATLAS, NIST AI RMF — and the EU AI Act obligations your product will be asked about.

Not included: model training, red-teaming of the foundation model itself, and remediation implementation.

Start with the scope your file is missing.

Sign the authorization, get the report in five business days.

Email usSee a sample