Skip to main content
AI coding agents such as Claude Code, Cursor and Codex run with the same access as the developer who starts them. Quint is a small endpoint on each developer machine plus a dashboard for your security team. The endpoint records what each agent said it was about to do and what the machine then observed it doing. The dashboard shows you that activity, names the risk in it, and lets you write rules about what agents may do.

Quickstart: enrol your first machine

Mint a token, install the endpoint on one Mac, and watch the machine and its first agent session appear in the dashboard.

​
What you can do with Quint

See what agents did

Every session, tool call, file change and network destination, per machine and per agent.

Investigate findings

A finding is one place where what an agent declared and what it did came apart, named by the risk it represents.

Set rules, measure first

Rules run in shadow mode until you arm them. You see what each rule would have done before it does anything.

​
Who uses it

  • Security and platform engineers open the dashboard daily to see what agents did, investigate findings and write rules.
  • Security leadership reads the Overview to see the state of the fleet at a glance.
  • IT administrators install the endpoint on machines, usually through an MDM.
  • Developers keep working as before. What Quint records and Quint is on my machine are written for them.
If a managed security provider or reseller runs Quint for you, or you are one, see the Partners tab.

​
Where it runs

Quint supports Apple Silicon Macs and records Claude Code, Codex, Cursor, Gemini CLI and other coding agents. Alerts go to Slack, Microsoft Teams or PagerDuty, and events to your SIEM. See Platform support and Integrations.