Independent security testing
Enterprise security for growing businesses.
Based in Austin, Texas, ZeroDark brings enterprise-grade security testing to startups and small and midsize businesses at prices they can budget for. We agree on scope and price up front, and you work directly with the tester who verifies your findings.
Reduced rates for organizations that serve our communities
We offer reduced rates to nonprofits, critical infrastructure operators, and healthcare organizations to help protect the services our communities depend on.
How we test and report
Your tester sets the boundaries, directs the work, and stands behind the findings.
Agree on the scope
Define what’s in scope, how we’ll test it, and what your auditor or engineering team needs from the report.
Test and verify
Your tester directs the work, combines hands-on checks with AI-assisted testing, and manually verifies findings.
How we use AIWalk through the report
Get a formal report with evidence, business impact, and prioritized fixes. Your tester explains what to address first.
Inspect a sample report
How we use AI
For external pentests, coordinated AI agent swarms help your tester cover at least three real-world attacker perspectives. Your tester verifies the findings and reports what each perspective revealed.
- Models run locally
- We run open-source and fine-tuned models locally on infrastructure we control.
- Runtime guardrails enforce the agreed scope
- Our custom harness coordinates our agents and applies multiple guardrails at runtime to limit testing to approved systems and activities. We run the agents inside NVIDIA OpenShell sandboxes to restrict file and network access.
- You control whether your data is used for training
- Training or fine-tuning our models with any client data requires your prior written permission.
How we handle client data
- Access to your data
- We use reasonable safeguards and limit access to people and service providers who need it to perform the work and are subject to confidentiality obligations. We make reasonable efforts to avoid accessing or retaining sensitive data beyond what the engagement requires.
- AI tools and source code
- We use Google Workspace for business operations. Using third-party cloud AI services or tooling in your engagement requires your prior written permission. Tell us during scoping if you have specific requirements for handling source code or evidence.
- Retention and return
- We retain data as reasonably necessary for the engagement and purposes such as legal obligations, backups and our retention policies. After the agreement ends, you can request return or destruction in writing. The agreement allows us to retain certain copies, which remain confidential.
- Sharing the report
- Reports are confidential. Under our agreement, you can share them with legal counsel, auditors, insurers and regulators who need access and meet the confidentiality terms. Ask for written consent before sharing with customers or other third parties.
Tell us what you need tested.
Share a rough scope and any deadline. We’ll help you choose an engagement.
