Built to watch. Not to snoop.
A security product has to earn trust twice — once by protecting you, and again by how it treats what it learns. Here is exactly how GreyArm works.
- Names, paths, signers and hashes of detected agents
- Which host each agent connected to, when and how often
- Names, addresses and vendors of devices on your network
- Alerts, your approvals and containment actions
- The content of prompts, replies or files
- Passwords, cookies or browsing history
- Keystrokes or screenshots
- Traffic from apps that aren’t talking to AI
Local first, by design.
Collection and analysis run inside the desktop sensor. It works fully offline and without an account.
Linking an account sends findings — not content — over TLS with a per-device token you can revoke at any time.
Monitoring needs no administrator rights. Containment asks for approval on the device, every time.
History is kept for your plan’s window, then deleted. Delete your account and your data goes with it.
Email verification, rate-limited sign-in, revocable sessions, and an audit log for every staff action.
Every detection rule is documented, so you can always see exactly why an alert fired.
Found something? Tell us.
Email support@greyarm.com with the subject “Security report”. Our Vulnerability Disclosure Policy covers scope, timelines and safe harbor for good-faith research.