Skip to main content

Your code stays yours, even with AI agents

"Enterprise-grade security" means nothing on its own. Here is exactly what protects your code, data and roadmap, what AI agents can and cannot reach, and how to verify all of it before you sign.

Request our SOC 2 report

We share the report under NDA. Tell us who you are and what your review needs; we reply within one business day.

By submitting you agree to our privacy policy.

Audited, not just claimed

  • SOC 2 Type IIIndependent audit of our security, availability and confidentiality controls.
  • ISO 27001Information security management compliant with ISO 27001.
  • GDPRPersonal data handled under GDPR, with data processing terms on request.
  • HIPAA readyDelivery practices prepared for healthcare workloads and BAAs.
  • AWS / Azure partnerCloud work follows provider-recommended security architectures.

Five controls on every engagement

NDA and IP assignment first
Every engagement starts with a confidentiality agreement and a clear assignment of intellectual property to you. Before any code is shared.
Your repositories, your accounts
Work happens in your repositories and cloud accounts, under your access controls. You keep ownership of the code, the environments and the secrets.
Least-privilege access
Engineers and AI agents get only the access a task needs. Access is removed when the engagement ends.
Encrypted communication
Project communication and file sharing run over encrypted channels.
AI tools you approve
You approve which AI coding tools touch your code before work starts. Business accounts only, never used to train models.
Every agent change is reviewed
A person reviews AI-generated code, then it is tested and shipped through CI like any other change. The history traces from commit to production.

Straight answers for your security review

Do AI coding agents see our proprietary code?

Only the code you give them access to. Agents work inside the repositories and environments you grant, under the same least-privilege rules as our engineers, and lose that access when the engagement ends. Every change they produce is reviewed by a person before it ships.

Which AI tools and models are used on our code?

Claude Code (Anthropic), OpenAI Codex, Cursor and GitHub Copilot. You approve the list before work starts and can restrict it to the tools your policies allow.

Is our code used to train AI models?

No. We use these tools only through business, team, enterprise or API accounts, whose terms exclude customer code from model training, with privacy modes turned on where a tool offers them. We never use personal or consumer plans on client code. If your policy requires it, agents can run under your own AI accounts or API keys instead of ours.

How long do AI vendors keep our code?

Code and prompts are handled under each vendor's standard business terms: they may be retained for a limited period for abuse and safety monitoring, and are not used for training. If you need different terms, agents can run under your own AI accounts, so your agreements with those vendors apply.

How do you keep secrets away from AI agents?

Agents are configured to ignore secret files such as .env files, keys and credentials, and secrets stay in a secrets manager, not in the repository. Engineers and agents get only the access a task needs, and it is removed when the engagement ends.

Which repositories can AI agents access?

The ones you grant, in your accounts, under your access controls. Engineers and agents get only the access a task needs. You can see and revoke that access at any time because it lives in your systems, not ours.

Who owns the code AI agents write for us?

You do. Every engagement starts with an NDA and a clear assignment of intellectual property to you, and the work lives in your repositories from the first commit.

Is AI-generated code reviewed before it ships?

Yes, every change. A senior engineer reviews it, then it is tested and shipped through CI like any other change, with a traceable history from commit to production.

Can we see your SOC 2 Type II report?

Yes. Use the request form on this page with your company and role and we will share the report under NDA, along with our ISO 27001 compliance documentation if you need it. We reply within one business day.

Will you complete our security questionnaire?

Yes. Send your vendor security questionnaire (SIG, CAIQ or your own format) and we will complete it as part of procurement. Questions about the AI tools in your engagement belong there too.

Before you go: the free guide

The Ultimate Guide to Agentic Product Development, as an instant PDF download.

No spam. See our Privacy Policy.