FAQ
Frequently asked questions
Straight answers on the model, your data, and what the agents do.
The model and your data
Does SecNode use the Memory Layer to train a model?
No. The Memory Layer is a persistent model of your environment that your agents read from and write to. It is context for reasoning about your company, and it is never used as training data.
About the Memory LayerDoes SecNode train on my code, findings, or evidence?
No. SecNode AI is not trained on your code, your environment, or anything the agents find there. What SecNode learns about your company stays in your Memory Layer and serves only you.
Is SecNode AI a wrapper around a third-party model?
No. SecNode AI is SecNode's own model, built for security reasoning, and it runs on SecNode's own EU-resident infrastructure. It cannot be switched off or changed by another provider's decision.
Where does my data live?
In the EU. SecNode is hosted and managed in the EU, with a dedicated single-tenant option.
Security & TrustCan SecNode run inside my own boundary?
Hosted in the EU today; deploy inside your own boundary on the enterprise tier.
Is SecNode GDPR-compliant?
Yes. Lawful processing, data-processing agreements, and data-subject workflows are handled by default, and the platform is DORA and NIS-2 aligned.
What the agents do
What does SecNode actually do?
It works like a security engineering hire. Seven specialist agents (web, API, and mobile pentesting, external attack surface, VDP triage, threat intel, and cloud security) run on SecNode AI and share your Memory Layer, and a remediation engineer in Slack and Microsoft Teams chases every confirmed fix to done.
Meet the agentsWhat access does SecNode need?
Read-only access to your code, cloud, and documentation, so the Memory Layer can be built and kept current. Testing runs only against targets inside a scope you declare.
How is a finding verified?
With a working reproduction. A finding only reaches your queue once an agent has demonstrated the issue is actually exploitable in your environment. A scanner's claim is not a finding.
Is it safe to run against production systems?
It has to be, or it is not deployable. Agents work inside explicit scope, use approval gates for intrusive actions, are non-destructive by default, and leave an audit trail for every step.
Does it replace the security team?
No. It gives the team leverage: agents do the investigation, testing, verification, and chasing, and your engineers review verified closures instead of raw queues. Risk decisions stay human.
What is out of scope?
Social engineering and physical testing. Denial-of-service and other volumetric testing are never run.
Who do I talk to?
A senior security engineer, not a sales rep. Tell us about your environment and what you want covered.
ContactPut senior-level security engineering on every system you own
See SecNode's agents run a live investigation against a real environment.