Try AI security agent swarms today

A pen test is a point-in-time assessment: one full engagement, one snapshot of your posture. Continuous penetration testing is always on: a swarm of AI security agents keeps testing your environment and learn it over time, catching what a single test would miss.

No. Agents plan attacks, run them, prove every finding with a working exploit, and generate the patch, all unattended. The only human decision is merging the fix.

A target is the domain or application scope you add for assessments, like app.example.com or api.example.com. Each subdomain or distinct application is one target.

Balanced is optimized for speed and frequent use (~6 hours). Deep provides more thorough analysis at moderate speed (~8 hours). Ultra is the most thorough option, best for occasional use (~12 hours).

Yes. Pen tests can be scheduled one-time, daily, weekly, or monthly. Continuous penetration testing runs by default, without manual scheduling.

Yes. On Starter, your first assessment is complimentary.

You can buy one-time credit packs with volume discounts, or enable usage-based billing so assessments keep running after prepaid credits are exhausted. Included plan credits refill each month.

Yes. Contact us for annual plans with discounted rates.

1 credit per finding. After you ship a fix, retest the finding to confirm it holds. Retests draw from the same credit pool as assessments.

Starter runs 1 assessment at a time. Scale runs up to 5 at once, so you can cover several targets in parallel. Enterprise concurrency is set by contract.

Yes, on every plan. Dual credential mode runs an assessment with two stored accounts, such as users in different tenants or a user and an admin, and tests whether one can reach the other's data or actions: IDOR, broken object-level authorization, role bypass, and cross-tenant access. Agents log in through SSO redirect flows and complete authenticator app (TOTP), SMS, and email MFA on their own.