Launching today

Covin

Launching today

Runtime security for autonomous AI agents

1 follower

Runtime security for autonomous AI agents

1 follower

Visit website

Covin is a security layer for autonomous AI agents. It sits between your agents and your systems, watches what they actually do, and blocks risky actions like secret exfiltration, destructive commands, and prompt-injection-driven behavior before they land. Covin Local is free and runs on your machine alongside your coding agents. Enterprise adds a team console, bait placement to catch rogue agents, and detection of multi-agent collusion.

Free

Launch tags:Developer Tools•Artificial Intelligence

Launch Team

Framer AI AgentsDesign and publish professional sites with AI

Promoted

Maker

📌

Hello Everyone! I'm Kartikey, maker of Covin. AI agents now run shell commands, read your files, and call APIs on their own. That's great until one gets prompt-injected, leaks a secret, or runs something destructive, and nothing sits in between to stop it. I built Covin to be that layer. It watches what agents actually do and blocks the risky stuff in real time. It started as a local guard for my own coding agents and grew from there: a free local version, plus an enterprise console for teams running fleets of agents. It's free to try at covin.dev. I'd love to hear which agent behaviors worry you most, and what you'd want Covin to catch next.

Report

4h ago